US10268880B2 - Device and method for recognizing animal's identity by using animal nose prints - Google Patents

Device and method for recognizing animal's identity by using animal nose prints Download PDF

Info

Publication number
US10268880B2
US10268880B2 US14/893,043 US201414893043A US10268880B2 US 10268880 B2 US10268880 B2 US 10268880B2 US 201414893043 A US201414893043 A US 201414893043A US 10268880 B2 US10268880 B2 US 10268880B2
Authority
US
United States
Prior art keywords
unit
image
nose
nose pattern
animal
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active, expires
Application number
US14/893,043
Other languages
English (en)
Other versions
US20160259970A1 (en
Inventor
Nam Sook Wee
Su Jin Choi
Haeng Moon Kim
Hyeong In Choi
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Iscilab Corp
Original Assignee
Iscilab Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Iscilab Corp filed Critical Iscilab Corp
Assigned to ISCILAB CORPORATION reassignment ISCILAB CORPORATION ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: CHOI, SU JIN, KIM, HAENG MOON, WEE, NAM SOOK
Publication of US20160259970A1 publication Critical patent/US20160259970A1/en
Priority to US16/185,354 priority Critical patent/US10552673B2/en
Application granted granted Critical
Publication of US10268880B2 publication Critical patent/US10268880B2/en
Active legal-status Critical Current
Adjusted expiration legal-status Critical

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01K—ANIMAL HUSBANDRY; AVICULTURE; APICULTURE; PISCICULTURE; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
    • A01K11/00—Marking of animals
    • A01K11/006—Automatic identification systems for animals, e.g. electronic devices, transponders for animals
    • G06K9/00362—
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01K—ANIMAL HUSBANDRY; AVICULTURE; APICULTURE; PISCICULTURE; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
    • A01K15/00—Devices for taming animals, e.g. nose-rings or hobbles; Devices for overturning animals in general; Training or exercising equipment; Covering boxes
    • A01K15/04—Devices for impeding movement; Devices for impeding passage through fencing, e.g. hobbles or the like; Anti-kicking devices
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01K—ANIMAL HUSBANDRY; AVICULTURE; APICULTURE; PISCICULTURE; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
    • A01K29/00—Other apparatus for animal husbandry
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/0059—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence
    • A61B5/0077—Devices for viewing the surface of the body, e.g. camera, magnifying lens
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/117—Identification of persons
    • A61B5/1171—Identification of persons based on the shapes or appearances of their bodies or parts thereof
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
    • A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
    • A61B5/6813—Specially adapted to be attached to a specific body part
    • A61B5/6814—Head
    • A61B5/6819—Nose
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/70—Means for positioning the patient in relation to the detecting, measuring or recording means
    • A61B5/702—Posture restraints
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G06K9/00885—
    • G06K9/4604—
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
    • G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/88—Image or video recognition using optical means, e.g. reference filters, holographic masks, frequency domain filters or spatial domain filters
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/168—Feature extraction; Face representation
    • G06V40/171—Local features and components; Facial parts ; Occluding parts, e.g. glasses; Geometrical relationships
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/172—Classification, e.g. identification
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B2503/00—Evaluating a particular growth phase or type of persons or animals
    • A61B2503/40—Animals
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B2562/00—Details of sensors; Constructional details of sensor housings or probes; Accessories for sensors
    • A61B2562/02—Details of sensors specially adapted for in-vivo measurements
    • A61B2562/0233—Special features of optical sensors or probes classified in A61B5/00
    • A61B2562/0242—Special features of optical sensors or probes classified in A61B5/00 for varying or adjusting the optical path length in the tissue
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B2576/00—Medical imaging apparatus involving image processing or analysis
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/0059—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/117—Identification of persons
    • A61B5/1171—Identification of persons based on the shapes or appearances of their bodies or parts thereof
    • A61B5/1176—Recognition of faces
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7253—Details of waveform analysis characterised by using transforms
    • A61B5/7257—Details of waveform analysis characterised by using transforms using Fourier transforms
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7253—Details of waveform analysis characterised by using transforms
    • A61B5/726—Details of waveform analysis characterised by using transforms using Wavelet transforms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00—ICT specially adapted for the handling or processing of medical images
    • G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing

Definitions

  • the present invention relates to an apparatus and method of an animal recognition using nose patterns, and particularly to an apparatus and method of an animal recognition comprising a body stabilizer unit to minimize movement of the subject animal for optimal nose pattern image acquisition, an image acquisition unit to obtain and store the nose pattern image, and an image recognition unit to generate, enroll, verify, and identify the raw or processed nose pattern code from the acquired nose pattern image.
  • Animal identification has been around for thousands of years as indicated in the Code of Hammurabi dating back to about 1754 BC.
  • the preferred method of marking a body part was used primarily to prevent the theft of valuable animal assets like horses.
  • animal identification serves an important role in the production management and disease control of livestock, the management of endangered and protected species, as well as the essential inspection process of animal imports and exports.
  • Globalization has increased the worldwide demand for animals for a variety of purposes, ranging from meat consumption to collecting exotic pets.
  • Electronic identification requires the use of an electronic ear tag, injectable transponder, or a ruminal blouse to contain and be scanned for the desired information.
  • unintentional damage to or intentional tampering of the microchip or antenna, as well as the unavailability of an appropriate scanning device can make identification impossible.
  • some studies have found that the material surrounding the microchip and antenna unit can cause tumors or tissue necrosis in the animals, providing significant reasons for concern among owners of companion and livestock animals.
  • Biometric identification relies on the intrinsic characteristics unique to individuals without the necessity of invasive procedures and, unlike the traditional methods or microchip identification, the biometric data of an animal cannot be doctored.
  • Current ongoing studies are seeking to make progress in animal iris and retina imaging, DNA analysis, and nose pattern imaging.
  • the first three have not been developed enough yet to be practically applicable in the field.
  • Korean Laid-open Patent Publication No. 10-2004-0008072 presents the technical configuration of portable information terminal for controlling cattle
  • Korea Laid-open Patent Publication No. 10-2004-0006822 discusses a method of remote bovine identification and health monitoring using previously scanned nose pattern data via the internet or a network.
  • the limitations in accuracy and the potential for aberration arising from human error during the process are prevalent.
  • the above methods cater only to bovine subjects and thus are inapplicable to animals with differently sized, shaped and patterned noses.
  • Korea Laid-open Patent Publication No. 10-2002-00066771 presents the technical configuration of a system of canine identification using nose pattern data through a communication network, but it does not specify the method of obtaining such data.
  • U.S. patent application Ser. No. 10/770,120 does disclose a technical construction of obtaining the nose pattern images of canine pets.
  • the pattern data are collected by either macro shooting with a specialized camera to compensate for the difficulty in focus adjustments while manually holding the subject animal's muzzle, or by getting an impression of the nose—similar to the traditional method—using some pliable plastic coated with ink or a paper or card comprising two different layers of chemicals in place of ink.
  • An object of the present invention is the acquisition and recognition of animal nose patterns without making direct physical contact with the nose.
  • Another object of the present invention is the acquisition of nose pattern images fit for recognition by utilizing a body stabilizer unit that minimizes obstructive movements in a subject animal that behaves uncooperatively out of fear or aggression toward the image acquisition equipment or the operator, and maintains the ideal frontal capturing angle on the subject's nose.
  • Yet another object of the present invention is the acquisition of good quality nose pattern images with the use of a body stabilizer unit designed to accommodate animals of different sizes and species.
  • Yet another object of the present invention is the acquisition of high quality nose pattern images by utilizing indirect illumination of appropriate wavelength regions applied through a light conduit subunit, light diffuser subunit, and spacer onto the subject's nose to prevent unwanted light reflections that may come off the layer of moisture on the nose surface of subject animals.
  • Yet another object of the present invention is the acquisition of high quality nose pattern images through the use of an image capture unit that is modifiable to accommodate subject animals of different species.
  • Yet another object of the present invention is to enable non-professional users to acquire nose pattern images fit for recognition with ease using the image acquisition unit.
  • Yet another object of the present invention is to make possible the identification of any animal with a discernible nose pattern, regardless of species- or breed-specific pattern types.
  • Yet another object of the present invention is to make identification possible regardless of the presence of extraneous physiological or environmental phenomena, such as moisture, hair, or dust of the subject animal's nose surface.
  • Yet another object of the present invention is to make identification possible despite reasonable variations in the image capturing angle.
  • Yet another object of the present invention is to generate a universal nose code irrelevant species or breed for use in identification.
  • Yet another object of the present invention is to use the most appropriate method of verification or identification for particular species or breeds.
  • Yet another object of the present invention is to increase the accuracy rate of recognition for each species of subject animals by comparing and matching previously stored nose pattern images to those newly obtained using the proper body stabilizer unit, image acquisition unit and image recognition unit.
  • a technical solution of the present invention is to provide an animal recognition apparatus comprising a body stabilizer unit, image acquisition unit, and image recognition unit.
  • Another technical solution of the present invention is to provide an animal recognition method comprising the following steps: selection of the appropriate body stabilizer for the species or breed of the subject animal, stabilization of the subject animal's body using the selected body stabilizer unit; acquisition of the nose pattern image by the image acquisition unit; storing of the acquired nose pattern image; generation of a nose pattern code from the acquired image; enrollment of the generated nose pattern code; and verification or identification of the subject animal by matching the newly obtained nose pattern code with previously enrolled nose codes.
  • Yet another technical solution of the present invention is to provide an animal recognition method comprising the following steps: acquisition of a nose pattern image using the body stabilizer unit and image acquisition unit; setting a region of interest (ROI) in the acquired nose pattern image, raw or processed; generation of a nose pattern code from the ROI or standardized ROI; enrollment of the newly generated nose pattern code; and verification or identification by determining the distance between the newly generated nose pattern code and previously enrolled nose codes.
  • ROI region of interest
  • the present invention has an advantageous effect in the fast and accurate recognition of animals through the acquisition of nose patterns without making direct physical contact with the nose.
  • Another effect of the present invention is the acquisition of nose pattern images fit for recognition by utilizing a body stabilizer unit that minimizes obstructive movements in a subject animal that behaves uncooperatively out of fear or aggression toward the image acquisition equipment or the operator, and maintains the ideal frontal capturing angle on the subject's nose.
  • Yet another effect of the present invention is the acquisition of good quality nose pattern images with the use of a body stabilizer unit designed to accommodate animals of different sizes and species.
  • Yet another effect of the present invention is the acquisition of high quality nose pattern images by utilizing indirect illumination of appropriate wavelength regions applied through a light conduit subunit, light diffuser subunit, and spacer onto the subject's nose to prevent unwanted light reflections that may come off the layer of moisture on the nose surface of subject animals.
  • Yet another effect of the present invention is the acquisition of high quality nose pattern images through the use of an image capture unit that is modifiable to accommodate subject animals of different species.
  • Yet another effect of the present invention is to enable non-professional users to acquire nose pattern images fit for recognition with ease using the image acquisition unit.
  • Yet another effect of the present invention is to make possible the identification of any animal with a discernible nose pattern, regardless of species- or breed-specific pattern types.
  • Yet another effect of the present invention is to make identification possible regardless of the presence of extraneous physiological or environmental phenomena, such as moisture, hair, or dust of the subject animal's nose surface.
  • Yet another effect of the present invention is to make identification possible despite reasonable variations in the image capturing angle.
  • Yet another effect of the present invention is to generate a universal nose code irrelevant species or breed for use in identification.
  • Yet another effect of the present invention is the use of the most appropriate method of verification or identification for the particular species or breed.
  • Yet another effect of the present invention is to increase the accuracy rate of recognition for each species of subject animals by comparing and matching previously stored nose pattern images to those newly obtained using the proper body stabilizer unit, image acquisition unit and image recognition unit.
  • FIG. 1 is a schematic diagram of the embodiment of the animal recognition apparatus described in the present invention.
  • FIG. 2 is a photograph of the nose pattern of a specific animal species (deer) to demonstrate an example of the subject of the present invention:
  • FIG. 3 is a photograph of the nose pattern of a specific animal species (dog) to demonstrate another example of the subject of the present invention.
  • FIG. 4 is a photograph of obstructive light reflections from the moisture naturally present on the surface of the nose of the subject animals.
  • FIG. 5 is a diagram illustrating the operation of the animal recognition apparatus in the present invention.
  • FIG. 6 is a presentation of the animal recognition apparatus in FIG. 5 shown from different angles.
  • FIG. 7 is a block diagram illustrating a configuration of the animal recognition apparatus in the present invention in which each of the parts, the body stabilizer unit, the image acquisition unit, and the image recognition unit are all separate;
  • FIG. 8 is a block diagram illustrating a configuration of the animal recognition apparatus in the present invention in which the body stabilizer unit and the image acquisition unit are separate from the image recognition unit;
  • FIG. 9 is a block diagram illustrating a configuration of the animal recognition apparatus in the present invention in which the image acquisition unit and the image recognition unit are separate from the body stabilizer unit;
  • FIG. 10 is a block diagram illustrating a configuration of the animal recognition apparatus in the present invention in which the body stabilizer unit, the image acquisition unit, and the image recognition unit are all connected.
  • FIG. 11 is a flowchart illustrating the method of operating the animal recognition apparatus in the present invention.
  • FIG. 12 is a block diagram schematically showing how to use the body stabilizer unit
  • FIG. 13 is a block diagram schematically showing how to use the posture stabilizer unit
  • FIG. 14 is a block diagram schematically showing how to use the position adjuster unit.
  • FIG. 15 is a diagram illustrating the application of the upper and lower body stabilizer units on two specific species (cow and deer);
  • FIG. 16 is a diagram illustrating the application of the upper and lower body stabilizer units on two other specific species (dog and cat).
  • FIG. 17 is a diagram illustrating the configuration of the head stabilizer unit.
  • FIG. 18 is a diagram illustrating the implementation of the appropriate posture stabilizer unit and position adjuster unit on a specific species (cow);
  • FIG. 19 is a diagram illustrating the implementation of the appropriate posture stabilizer unit and position adjuster unit on another species (deer);
  • FIG. 20 is a diagram illustrating the implementation of the appropriate posture stabilizer unit and position adjuster unit on two small species (dog and cat).
  • FIG. 21 is a flowchart illustrating a method of operating the body stabilizer unit in the present invention.
  • FIG. 22 is a block diagram illustrating the configuration of the image acquisition unit.
  • FIG. 23 is a block diagram schematically illustrating the image capture unit and the image analysis unit within the image acquisition unit.
  • FIG. 24 is a block diagram illustrating the configuration of the image capture unit that moves the lens module and sensor according to the distance adjustment principle of the distance adjuster module.
  • FIG. 25 is a block diagram illustrating the configuration of the image capture unit that adjusts the distance between the lenses in the lens module according to the distance adjustment principle of the distance adjuster module.
  • FIG. 26 is a diagram illustrating the configuration of the front unit of the image capture unit.
  • FIG. 27 is a diagram illustrating the configuration of the illumination unit.
  • FIG. 28 is a block diagram illustrating the method of obtaining nose pattern images that are usable by the image recognition unit through the image capture unit.
  • FIG. 29 is a diagram illustrating the method of adjusting the field of view and focus by moving the lens module or sensor in the image capture unit;
  • FIG. 30 is a diagram illustrating the method of adjusting the field of view and focus by moving the lenses within the lens module of the image capture unit.
  • FIG. 31 is a diagram illustrating how to manipulate the field of view adjuster lens, the length of the spacer, and the type of front unit to fit the noses of larger subject animals (cow and deer);
  • FIG. 32 is a diagram illustrating how to manipulate the field of view adjuster lens, the length of the spacer, and the type of front unit to fit the noses of medium-sized subject animals (dog);
  • FIG. 33 is a diagram illustrating how to manipulate the field of view adjuster lens, the length of the spacer, and the type of front unit to fit the noses of smaller subject animals (cat or very small dog).
  • FIG. 34 is a set of photographs showing the results of using the three different types of the front unit.
  • FIG. 35 is a pair of photographs comparing the results of using direct illumination of a conventional camera and the indirect illumination of the present invention to acquire the nose pattern image of the same individual.
  • FIG. 36 is a block diagram schematically describing the image analysis unit of the present invention.
  • FIG. 37 is a diagram illustrating the method of nose pattern image acquisition during capture mode.
  • FIG. 38 is a flowchart illustrating the method of nose pattern image acquisition by the image acquisition unit.
  • FIG. 39 is a block diagram schematically describing the image recognition unit of the present invention.
  • FIG. 40 is a flowchart illustrating the method of analyzing and recognizing nose pattern images.
  • FIG. 41 is a block diagram schematically describing the region of interest fixing unit.
  • FIG. 42 is a diagram illustrating the method of finding the boundary of the nostril.
  • FIG. 43 is a pair of diagrams illustrating the method of approximating the boundary of the nostrils with curves (circle and ellipse, respectively)
  • FIG. 44 is a diagram illustrating the method of obtaining the region on the opposite side of each nostril, that is located in the exterior of the approximated curves (circle/ellipse).
  • FIG. 45 is a diagram illustrating the method of selecting the rectangular area between the approximation curves (circle/ellipse) as the region of interest.
  • FIG. 46 is a diagram illustrating the differences in the region of interest resulting from using circular and elliptical approximation curves on the same nose pattern image.
  • FIG. 47 is a diagram illustrating the process of generating a standardized region of interest from a previously fixed region of interest.
  • FIG. 48 is a simplified block diagram describing the nose pattern code generation unit.
  • FIG. 49 is a block diagram illustrating the process of generating nose pattern codes from the region of interest.
  • FIG. 50 is a diagram illustrating how to divide the region of interest into smaller cell blocks of specified dimensions, from which frequency transform codes are generated.
  • FIG. 51 is a diagram illustrating the comparison of a theoretical calculation area and the actual area when generating the frequency transform code using Gabor transform, Gabor cosine transform, Gabor sine transform, etc.
  • FIG. 52 is a simplified block diagram describing the nose pattern code matching unit.
  • FIG. 53 is a diagram illustrating the method of nose pattern code identification through simple matching.
  • FIG. 54 is a diagram illustrating a situation in which different regions of interest have been selected from the same nose pattern image of the same individual for matching.
  • FIG. 55 is a diagram illustrating the method of nose pattern code identification through shift matching.
  • FIG. 56 is a diagram illustrating a matching situation in which the regions of interest of the nose pattern code selected from the same individual have nonidentical vertical and horizontal proportions.
  • FIG. 57 is a diagram illustrating the method of nose pattern code identification through block-wise shift matching.
  • FIG. 58 is a diagram illustrating the process of nose pattern code identification through shift matching using Gabor sine transform.
  • FIG. 59 is a diagram illustrating the process of nose pattern code identification through block-wise shift matching using Gabor sine transform.
  • FIG. 60 is a diagram illustrating the method of nose pattern code identification (one-to-many matching).
  • Verification refers to one-to-one (1:1) matching
  • Identification or Searching refers to one-to-many (1:n) matching
  • Recognition encompasses both the Verification and Identification processes.
  • nose pattern relates to how the beads and grooves form geometric patterns on the nose surface, and it should be noted that the size and intricacy of the patterning can vary even within the same species.
  • the present invention describes an animal recognition method and apparatus for animals with unique nose patterns (subject animals) through the acquisition of identifiable nose pattern images by utilizing a body stabilizer unit to minimize movement and resistance in the subjects; an image acquisition unit to capture said images; and an image recognition unit that generates processed nose pattern images via noise reduction and image quality reinforcement techniques, and from it, nose pattern codes for enrollment and identification.
  • the image acquisition unit 102 may include the image analysis unit to be described later; or, the image analysis unit may be included in the image recognition unit 103 .
  • the image analysis unit may be included in the image recognition unit 103 .
  • various configurations and modifications are entirely possible to suit the user's request or the designer's purpose.
  • FIG. 2 and FIG. 3 show two examples of nose patterns, as taken using a model of the image acquisition unit.
  • the key features are the nostrils and the beading and grooving patterns, where beads are areas of raised nose skin and grooves are the narrow valleys surrounding each bead.
  • FIG. 3 shows the nose pattern on a dog where, while the specific size and shapes differ, a similar beading and grooving phenomenon can be found.
  • the beads tend also to be relatively larger, while in smaller species like cats and dogs the beads tend to be proportionately smaller.
  • the size of the nose area as generally dependent on the body size can affect the size of the beading patterns; and so it is necessary that the recognition apparatus and method take into consideration the patterning variations in different species and breeds.
  • a healthy nose maintains a layer of moisture on the surface, which aggressively reflects light in photos taken under natural settings. This also adversely affects the results of any method that relies on contact imprinting, such as with paper or glass, as the moisture can often cause blurring and image distortions. As for image capturing, there is also ample possibility that the moisture would reflect light or absorb the infrared range. Thus, the moisture presents an unavoidable problem that needs to be addressed and solved.
  • the temperament of the subject animal is a factor as well, as it can vary from one individual to another even in the same species or breed. While some subjects are naturally tame and cooperative, others are more timid or aggressive especially towards the various (capturing, stabilizing, or illumination) apparati and human operators, making the work out in the field difficult or unsafe, especially for a non-professional user. Therefore, an effective method and apparatus must not aggravate the subject animals.
  • the technical configuration of the animal recognition apparatus is as follows: body stabilizer unit, image acquisition unit, and image recognition unit.
  • the body stabilizer unit refers to the set of devices that prevent disruptive movements from the subject animal;
  • the image acquisition unit refers to the software and hardware described in the present invention necessary to acquire the nose pattern images of a variety of animals; and
  • the image recognition unit refers to the software and hardware needed for nose pattern image recognition.
  • FIGS. 5 and 6 illustrate a practical application and operation of the overall animal recognition apparatus in the present invention
  • the block diagrams in FIGS. 7, 8, and 9 show the connective combinations among the body stabilizer, image acquisition, and image recognition units.
  • the animal recognition apparatus may be configured with a certain level of flexibility depending on the given animal or setting, where all three component units could be connected, or just two, or all three are set up separately.
  • the flowchart in FIG. 11 summarizes the method of animal recognition in the present invention, starting with S 1101 selecting and S 1102 fitting the animal into the appropriate body stabilizer unit; S 1103 fixing the nose of the subject onto the image acquisition unit and S 1104 acquiring the nose pattern image; S 1105 at the image recognition unit, generating a nose pattern code using the raw or processed nose pattern image, and S 1106 enrolling and identifying the individual using the nose pattern code.
  • S 1101 selecting and S 1102 fitting the animal into the appropriate body stabilizer unit
  • S 1103 fixing the nose of the subject onto the image acquisition unit and S 1104 acquiring the nose pattern image
  • S 1105 at the image recognition unit generating a nose pattern code using the raw or processed nose pattern image
  • S 1106 enrolling and identifying the individual using the nose pattern code.
  • the purpose of the body stabilizer unit is to temporarily control the movement or resistance of the subject animal in reaction to the illumination or the operator, such as head turning or aggressive behavior, during the nose pattern image acquisition process to yield the best quality image. This is a safety measure against the animal accidentally harming itself or the human operators, which would incur added difficulties as well as cost. Moreover, an ideal nose pattern image is one taken from head-on and this is difficult to obtain with a highly uncooperative subject animal without the help of a body stabilizer.
  • the four primary functions of the body stabilizer unit are as follows: minimize the motion of the subject animal during the image acquisition process, act as a safety measure to protect the operator and apparatus, protect the subject animal from self-harm, and hold the nose in place for the best angle of image capture.
  • the body stabilizer unit comprises the posture stabilizer unit 1201 , and also the position adjuster unit 1202 to accommodate the subject's stance width.
  • FIG. 13 shows a more detailed breakdown of the posture stabilizer unit.
  • the subject animal can and often will resist with the whole body and thereby cause blurry images.
  • This problem may be mitigated by stabilizing the neck and shoulder area (upper body), as well as the back, front, and hind legs (lower body). Vets commonly forgo the usage of anesthesia during procedures whenever possible by applying pressure on the nape or shoulder of the patient animals; the body stabilizer is meant to simulate this method by allowing the subject animal to rest its head on the chin support while holding it in position with the head stabilizer unit 1302 and applying appropriate pressure using the upper body stabilizer unit 1301 . Further movement in the lower body, especially in the case of larger animals whose powerful legs may pose a danger to the operators and equipment, may additionally be held in check by the lower body stabilizer unit 1303 .
  • the configuration of the position adjuster unit is modifiable in accordance with the posture stabilizer unit settings, as the operator sees fit for the subject animal. Possible additions are the height adjuster unit 1404 to the upper body stabilizer unit 1401 ; stance width adjuster unit 1406 to the lower body stabilizer unit 1403 ; and the horizontal balance adjuster unit 1405 and the height adjuster unit 1407 to the head stabilizer unit 1402 .
  • FIGS. 15 and 16 each show an example of the posture stabilizer unit with upper and lower body stabilizer units as appropriate for larger animals as cows and deer, and for smaller animals as dogs and cats, respectively.
  • the upper and lower body stabilizers may be set up in various combinations—each independently, in conjunction, or connected at certain parts.
  • the upper body stabilizer unit comprises the upper body stabilizing brace subunit 1501 , 1601 and the upper body stabilizing brace lock subunit 1502 , 1602 ; the upper body stabilizing brace pressure adjuster subunit 1503 , 1603 is optional.
  • the upper body stabilizing brace subunit 1501 , 1601 may be made into a cover type, with durable synthetic fabrics, or with length-adjustable belts or cables.
  • the upper body stabilizing brace lock subunit 1502 , 1602 prevents the brace subunit from coming undone during the procedure, and may be manual or electronic.
  • the upper body stabilizing brace pressure adjuster subunit 1503 , 1603 allows the upper body stabilizing brace subunit to apply pressure on the subject animal by, for example, inflating the brace with some gas or liquid with the use of a pressure injector paired with a pressure monitor subunit.
  • the lower body stabilizer unit comprises the lower body stabilizing brace subunit 1504 , 1604 and the lower body stabilizing brace lock subunit 1505 , 1605 ; the lower body stabilizing brace pressure adjuster subunit 1506 , 1606 , as well as the lower body stabilizer supporting subunit 1507 , 1607 are optional.
  • the lower body stabilizing brace subunit 1504 , 1604 may be made into a cover type, with durable synthetic fabrics, or with length-adjustable belts or cables.
  • the lower body stabilizing brace lock subunit 1505 , 1605 prevents the brace subunit from coming undone during the procedure, and may be manual or electronic.
  • the lower body stabilizing brace pressure adjuster subunit 1506 , 1606 allows the lower body stabilizing brace subunit to apply pressure on the subject animal by, for example, inflating the brace with some gas or liquid with the use of a pressure injector pair with a pressure monitor subunit.
  • the lower body stabilizer supporting subunit 1507 , 1607 fastens the lower body stabilizer unit to the ground or at a certain distance from the equipment, and may be made up the lower body supporting subunit and the lower body supporter connector subunit.
  • the lower body supporting subunit and the lower body supporter connector subunit may take many forms to suit the subject animal, and may be made of steel or other metals, as well as durable synthetic fibers, rubber, or fabric.
  • the head stabilizer unit in FIG. 17 comprises the chin support 1701 , and the stabilizing muzzle subunit 1702 that holds the subject's nose in the correct position.
  • the chin support 1701 may be made of various materials as wood, plastic, rubber, or metal, and should withstand the weight of the subject's head while providing a comfortable headrest and room for additional supporting attachments.
  • the stabilizing muzzle subunit 1702 will be used when the head movement cannot be controlled with the chin support alone, and may be made into a cover type with durable synthetic fabrics, or with length-adjustable belts or cables, to span the muzzle area.
  • the stabilizing muzzle lock subunit 1703 prevents the brace subunit from coming undone during the procedure, and may be manual or electronic.
  • the stabilizing muzzle pressure adjuster subunit 1704 allows the stabilizing muzzle subunit to apply pressure on the subject animal's muzzle by, for example, inflating the brace with some gas or liquid with the use of a pressure injector paired with a pressure monitor subunit.
  • the head stabilizer unit may also have a head stabilizer support 1705 that fastens the head stabilizer unit to the ground or at a certain distance from the equipment while supporting the weight of the subject animal's head, and may be made into various shapes using durable materials as wood, stone, or metal.
  • the position adjuster unit adjusts the settings of the posture stabilizer unit as per each animal's physical characteristics in order to produce the most comfortable position for the subject, and comprises the height adjuster unit, horizontal balance adjuster unit, and stance width adjuster unit.
  • the examples in FIGS. 18 and 19 show the subject animal (cow and deer, respectively) fitted into the upper and lower body stabilizer units where the height adjuster units 1801 , 1804 , 1901 , 1904 are set to accommodate the subject's height.
  • the height adjuster 1801 , 1901 connects the upper and lower body stabilizer units and may comprise the connector subunit 1802 , 1902 made with belts or cables, and the length adjuster subunit 1803 , 1903 that adjusts the length of the connector subunit.
  • the height adjuster unit 1804 , 1904 for the head stabilizer unit may also comprise the connector subunit 1805 , 1905 connecting the chin support to the ground and the chin support height adjuster unit 1806 , 1906 .
  • the horizontal balance adjuster unit 1807 , 1907 placed inside or outside the chin support and comprising a horizontal balance sensor with a display monitor, positions the chin support under the subject animal's head to directly face the image acquisition unit.
  • the horizontal balance sensor may comprise gravity, gyro, or pressure sensors.
  • the stance width adjuster unit 1808 , 1908 may be used when the lower body is fastened in the lower body stabilizer unit, and may comprise connector subunit 1809 , 1909 of belts or cables and a length adjuster subunit 1810 , 1910 that connects both sides of the lower body.
  • the height adjuster unit 2001 that adjusts the height of the head stabilizer unit may comprise a connector subunit 2002 that connects the chin support to the ground and a height adjuster subunit 2003 that adjusts the height of the connector subunit.
  • the horizontal balance adjuster unit 2004 placed inside or outside the chin support and comprising a horizontal balance sensor with a display monitor, positions the chin support under the subject animal's head to directly face the image acquisition unit.
  • the horizontal balance sensor may comprise gravity, gyro, or pressure sensors.
  • the stance width adjuster unit 2005 may be used when the lower body is fastened in the lower body stabilizer unit, and may comprise a connector subunit 2006 of belts or cables and a length adjuster subunit 2007 that connects both sides of the lower body.
  • the sequence of operation for the body stabilizer unit is as follows: S 2101 select the appropriate body stabilizer unit for the subject animal by taking into consideration the overall size, leg length, feet size, head size, and the relative location of the nose; S 2102 fit the subject animal into the upper body stabilizer unit; S 2103 fasten the upper body by utilizing the upper body stabilizing brace subunit and upper body stabilizing brace pressure adjuster subunit to fit the shoulder width; S 2104 fit the subject animal into the lower body stabilizer; S 2105 fasten the lower body by utilizing the lower body stabilizing brace subunit and lower body stabilizing brace pressure adjuster subunit to fit the ankles or legs; S 2106 set the stance width adjuster, and also the height adjuster to fit the subject's height if necessary to connect the upper and lower body stabilizer units; S 2107 fasten the head by utilizing the head stabilizer unit, making sure to set the height adjuster unit to the correct height and the horizontal balance adjuster unit to have the nose facing the image acquisition unit head-on. This sequence of events may be modified
  • the purpose of the image acquisition unit is the capture and acquisition of nose pattern images. This can seem conceptually innocuous but the execution of it is nothing but, due to the morphological diversity of the nose and nose patterns, as well as the physiological nature of the rhinarium that yields unwanted light reflections.
  • the six primary functions of the image acquisition unit are as follows: acquire good quality nose images usable by the image recognition unit without relying on the traditional methods that mandate direct contact; acquire good quality nose images from a wide variety of species; not be affected by a subject animal's particular size, shape, or physiology; employ a special kind of illumination to avoid issues with light reflections from the wet nose; and enable non-professional users to achieve the above five with ease.
  • the image acquisition unit comprises the image capture unit 2201 that photographically captures nose pattern images, and also possibly the image analysis unit 2202 that analyzes the captured images and processes certain signals and information.
  • the image capture unit comprises the capture unit 2301 , the front unit 2302 that adjusts the field of view (FOV) and capture distance for each subject while blocking out the ambient light for a more controlled environment, and additionally the illumination unit 2303 that provides indirect illumination to overcome the light reflection issue ( FIG. 23 ).
  • the “acquisition” of nose pattern images refers to the entirety of the process of capturing and storing of nose pattern images.
  • “acquiring” nose pattern images includes capturing photographic images by the image sensor in the image capture unit and storing the captured images in the buffer for further actions, including but not limited to image processing, best image selection, etc.
  • the image capture unit may comprise the lens module 2401 with two or more lenses; an image sensor 2402 (CMOS or CCD); and the distance adjuster module 2403 that controls the FOV and focus by moving the lens module and sensor, thereby manipulating the distances between the lenses and between the lens module and the sensor within the front unit 2404 .
  • the distance adjuster module 2401 moves the lens module or a plurality of lenses, and comprises a small motor and rack gear that the converts motor's circular motion to linear motion.
  • a guide rail that allows the lens module 2401 and sensor 2402 , in linear periodic motion by the rack gear, to move between predetermined positions may also be installed.
  • the image capture unit may also comprise the lens module 2501 and sensor 2502 in fixed positions, with the distance adjuster module 2503 only controlling the distances between the lenses within the lens module ( FIG. 25 ).
  • the front unit may comprise a front cover 2601 that surrounds and/or comes in contact with the skin around the nose when the nose enters the front unit; a FOV adjuster lens 2603 ; a spacer 2602 that adjusts the distance between the subject's nose and the FOV adjuster lens 2603 .
  • the front cover 2601 and spacer 2602 may come in variable shapes or sizes to accommodate different species or breeds.
  • the front cover 2601 should be of a color that is best suited for blocking out ambient light, most likely black or other dark hues, and made of materials that do not agitate the subject animals, such as synthetic fibers, rubber, textile, or plastic.
  • the front cover also may be imbued with a calming scent for the subject animals, and made to be detachable for easy substitution when dealing with subjects of different physical requirements during the same session.
  • the standard FOV adjuster lens 2603 is modeled after the nose size of a typical (medium sized) dog; a reducing lens is used instead for larger noses, and a magnifying lens for smaller noses.
  • the standard lens refers to a single or a set of lenses that allows the framing of a typical dog nose, and the reducing and magnifying lenses are made in relation to the standard.
  • the spacer 2602 consists of the exterior that the blocks the light coming from the outside, and the interior that surrounds the nose of the subject animal, and possibly also houses an illumination unit.
  • the length of the spacer which determines the distance between the FOV adjuster lens and the subject's nose, may be optimized using field trial results. It also may be efficient to have pre-designed, detachable front units with spacers and FOV adjuster lenses set to fit particular species or nose sizes based on experimental results.
  • the illumination unit in FIG. 27 seeks to eliminate the issues that arise from the reflection and absorption of light by the moisture on the nose surface by incorporating a light source 2701 of a specific wavelength region (that poses no threat to the health of the subject animal) in indirect illumination, wherein the light travels through a light conduit subunit 2703 and light diffuser subunit 2702 .
  • the light source 2701 , light diffuser subunit 2702 , and light conduit subunit 2703 may vary to suit different species of subject animals.
  • the light source 2701 should have adjustable luminosity, avoid the infrared region that can be absorbed by the moisture and the UV region that can cause tissue damage, and be optimized to suit the particular characteristics of a species. Any type of light source consistent with the above description would suffice.
  • the light diffuser subunit 2702 partially absorbs and reflects light from the light source through the diffuser surface to indirectly illuminate the whole nose surface inserted into the front unit.
  • the amount of light that eventually passes through the diffuser may be controlled with the type of material used, such as Hanji (traditional Korean paper handmade from mulberry trees), translucent tracing paper, or a special type of glass, and similar material may also be used to line the interior of the light conduit subunit 2703 .
  • the image capture unit 2801 employs a variety of ways to obtain good quality images.
  • the FOV and focus are controlled by the distance adjuster module 2805 either by moving the lens module 2811 A or sensor 2812 , or by moving the plurality of lenses 2813 within the lens module while the lens module 2811 B and sensor 2812 stay fixed.
  • the front unit 2803 adjusts the FOV with the FOV adjuster lens 2806 , and the focus by changing distance between the FOV lens 2806 and lens module 2815 via variable spacer 2807 length.
  • the illumination unit 2804 employs a light source 2808 of a wavelength region optimal for nose images and the light conduit subunit 2809 and light diffuser subunit 2810 for indirect illumination, which is essential to producing good quality images without obstructive reflections ( FIG. 28 ).
  • the first method involves moving the lens module 2901 or the sensor 2902 , independently or concurrently, along a linear axis using the distance adjuster module.
  • the change in the position of the lens module 2901 or sensor 2902 changes the distance (a) between the two, and the distance (b) between the lens module 2901 and the FOV adjuster lens within the front unit 2903 , thereby changing the FOV and focus.
  • the length of the spacer 2903 is preset for the particular subject animal, and the distances (d 1 , d 2 ) between the lenses 2904 in the lens module are also fixed.
  • the values of a and b could also be set in advance for specific species so that non-professional users could easily carry out the capture process.
  • the second method involves moving the lenses 3004 within the lens module along a linear axis, thereby changing the distances between the lenses (d 1 , d 2 ), to change the FOV and focus.
  • the length (c) of the front unit 3003 is set in advance for the appropriate species and therefore a fixed value; and the distance (a) between the lens module 3001 and the sensor 3002 , and the distance (b) between the lens module and the FOV adjuster lens in the front unit 3003 are also fixed.
  • the distance adjuster module may be configured to move with the lenses 3004 within the lens module so that only d 1 and d 2 values can be manipulated.
  • the values of a and b could also be set in advance for specific species so that non-professional users could easily carry out the capture process.
  • the front unit uses the FOV adjuster lens and the length of the spacer to manipulate the FOV and focus.
  • FIGS. 31, 32, and 33 illustrate the how different combinations of the FOV adjuster lens and spacer length may be used to accommodate subject animals of different sizes.
  • the standard FOV adjuster lens is modeled after the nose size of a typical (medium sized) dog; a reducing lens is used instead for larger noses, and a magnifying lens for smaller noses.
  • the standard lens refers to a single or a set of lenses that allows the framing of a typical dog nose, and the reducing and magnifying lenses are made in relation to the standard.
  • the length of the spacer may be changed, depending on the nose size of the subject animal, to adjust the focus (distance).
  • the FOV adjuster lens 3101 should be a reducing lens, and the spacer 3102 should be set in advance to a length appropriate to get the right focus on the subject animal. Also, since the length (c) of the spacer 3102 can change, the distance (a) between the lens module 3103 and the sensor 3104 and the distance (b) between the lens module and the FOV adjuster lens in the front unit 3101 may also change.
  • a standard FOV adjuster lens 3201 should be used for medium-sized noses, paired with an appropriately lengthened spacer 3202 . Also, since the length (c′′) of the spacer can change, the distance (a′′) between the lens module 3203 and the sensor 3204 and the distance (b′′) between the lens module and the FOV adjuster lens in the front unit 3201 may also change.
  • the FOV adjuster lens 3301 should be a magnifying lens, and the spacer 3302 should be set in advance to a length appropriate to get the right focus on the subject animal. Also, since the length (c′′′′) of the spacer can change, the distance (a′′′′) between the lens module 3303 and the sensor 3304 and the distance (b′′′′) between the lens module and the FOV adjuster lens in the front unit 3301 may also change.
  • FIG. 34 shows the results of using the different front unit settings on a dog, demonstrating the importance of choosing the right one; the first image was taken with the magnifying setting, the second with the standard, and the third with the reducing.
  • FIG. 35 shows a side-by-side comparison of the same dog nose image captured using the conventional direct illumination (camera flash) and the indirect illumination of the illumination unit.
  • the illumination unit controls the light reflections—which appear as white flecks on the left image, and are highly obstructive to accurate identification—from the moisture on the nose surface by achieving indirect illumination through the use of a special light source, light conduit subunit, and light diffuser subunit.
  • the light source should avoid the infrared region that can be absorbed by the moisture and the UV region that can cause tissue damage, and be optimized to suit the particular characteristics of a species.
  • FIG. 35 demonstrates that using the conventional camera flash does not help to contrast the nostril from the nose surface area, while indirect illumination results in clear boundary distinctions. This affects the ease with which the nostril boundary can be established, and thus indirectly illuminated images will generally increase the recognition accuracy.
  • the image analysis unit analyzes the nose pattern images acquired by the image capture unit, manages various kinds of information and signals from the process, and may also be attached not only to the image acquisition unit but also the image recognition unit.
  • the image analysis unit may comprise the main processor unit 3601 , the buffer 3602 , the database (DB) 3603 , and the communication unit 3604 .
  • DB database
  • a display unit 3605 may be added so that the operator may see the images captured by the image capture unit in real time, and select and acquire good quality images.
  • the main processor unit selects nose pattern images that are of sufficient quality to be used in the image recognition unit, out of all the images captured by the image capture unit.
  • each image is given individual scores on specific variables, and images that pass the threshold set by the image analysis unit are selected. If none out of a particular batch meet the threshold, then that whole group is discarded and a request for a new batch is sent to the image capture unit.
  • the images are evaluated on such criteria as, the amount of light reflection, sharpness, contrast ratio, ROI for capture, noise level, etc; and only those images that pass the threshold for each variable are accepted.
  • the threshold the one with the highest total score (sum of individual scores) is selected, and this process may take place simultaneously as the image acquisition in the image analysis unit or the image recognition unit.
  • variables there are two types of variables: those that are not related to species-specific characteristics (A 1 -A 3 ) and those that are (A 4 -A 12 ).
  • the former includes sharpness A 1 , contrast A 2 , and noise level A 3 ;
  • the latter includes ROI for capture A 4 , presence of light reflection A 5 , nostril location A 6 , sharpness of nostril image A 7 , contrast level of nostril image A 5 , noise level of nostril image A 9 , sharpness of the border between the nostril and ROI A 10 , contrast level at the border between the nostril and ROI A 11 , and noise level at the border between the nostril and ROI A 12 .
  • Variables may be appropriately added to or subtracted from the above list depending on a subject animal species' particular characteristics (Table 1).
  • the above total score is the weighted sum of the individual scores, and therefore the degree of importance of a particular variable may be reflected by adjusting the weight value.
  • the image acquisition unit can capture a batch of nose pattern images.
  • the image capture unit is not on standby in sleep mode, it is in automatic or manual capture mode.
  • Automatic capture mode receives the threshold values for the species or breed from the DB, and compares them to the individual scores of the captured images at the main processor unit.
  • manual mode the user operates the image acquisition unit, visually makes an estimated evaluation of the individual scores of each variable and makes the capturing decision if these scores are deemed satisfactory.
  • the sleep mode is a standby mode before entering the capture (recording) mode, and the capture mode is for the final nose pattern image acquisition.
  • the image acquisition unit may transition from sleep mode to capture mode when the user presses a designated button on the display unit.
  • FIG. 37 illustrates the acquisition of a batch of nose pattern images in capture mode.
  • the lens module or sensor in the capture unit moves into position according to the preset value.
  • the transition from capture mode to sleep mode occurs when the best (threshold-passing) image is successfully selected by the main processor unit from among the batch saved to the buffer. If the time at the start of the recording is T_start and the end of the recording is T_end, then a n number of images are acquired during that time at a constant rate per second.
  • the per second frame rate will vary depending on the resolution, and may also vary depending on the hardware configuration and the type of camera.
  • the main processor unit may also alert the operator of the end of the capture process through a notification on the display unit.
  • the minimum hardware components of the main processor unit are the CPU, RAM, and nonvolatile memory (ROM, flash memory, etc).
  • the CPU performs all of the operations carried out by the image analysis unit.
  • the nonvolatile memory is mounted the resident program where the threshold values are stored, the individual scores of nose pattern images are evaluated, and the algorithm that enables the saving of the selected images and all related information to the buffer is stored.
  • nonvolatile memory may not be efficient for speedy processing, in which case RAM may be a useful addition to the main processor unit.
  • the buffer stores a variety of information that arise while the main processor unit is in the process of selecting threshold-passing images, and may consist of a flash memory or a DB. Since the DB on the buffer can be changed any time by the user, the DB of the buffer generated by the image analysis unit should preferably be stored in the flash memory.
  • the parameter DB stores the threshold values and individual scores selected by the main processor unit in the image acquisition unit.
  • the communication unit relays information between the image capture unit and image analysis unit.
  • the communication unit is tasked with the output of signals for positioning commands during capture and alerting the user of mode changes, and is thus basically equipped with a signal transmitter for outputting instruction signals.
  • the signal transmitter may comprise one or more of the following: audible signal generator (for voice or other sounds), visual signal generator (for LED or flash), and vibration generator.
  • a display unit possibly comprising a mirror or LCD may be supplemented to enable a quick and easy review of the images obtained by the image capture unit.
  • FIG. 38 An example of the image acquisition process is illustrated in FIG. 38 ; the order of events need not be limited as follows.
  • S 3801 the operator selects the species on the display unit to start the automatic mode, or chooses the manual mode;
  • S 3802 in automatic mode once the species selection is made, pressing the capture button starts the acquisition of the batch of nose pattern images at n frames per second while the lens module is shifted about within the preset range of positions (adjusting values of a and b).
  • the image capture unit, illumination unit and front unit are automatically adjusted to accommodate the subject animal based on the values stored in the parameter DB for FOV, focus, luminosity, etc.
  • manual mode the operator visually evaluates the features and variables of the nose pattern images through the display unit, and selects the best image.
  • the image recognition unit generates processed nose pattern images and nose pattern codes for enrollment and identification.
  • the image recognition unit should be capable of the following: identify any individual animal with viable nose patterns, regardless of idiosyncratic characteristics; identify regardless of the extraneous physiological phenomena (such as moisture or hair, etc); compensate for certain distortions that occur in images captured from different angles and perform accurate identification; create universal nose pattern codes for identification for any species or breed with viable nose patterns; and employ the best method of identification when performing matching within a known species or breed.
  • the image recognition unit may comprise the region of interest (ROI) fixing unit 3901 , the nose pattern code generation unit 3902 that generates the nose pattern code from the fixed ROI, the nose pattern code matching unit 3903 , and the nose pattern code database (DB) 3904 where the generated nose pattern codes are stored during the enrollment and identification stage.
  • ROI region of interest
  • DB nose pattern code database
  • Possible additions are the image processing unit 3905 , which processes the nose pattern image, if necessary, before setting the ROI; and the standardized ROI fixing unit 3906 , which standardizes the ROI before setting the ROI and generating the nose pattern code.
  • the aforementioned image analysis unit of the image acquisition unit may be configured into the image recognition unit.
  • FIG. 40 illustrates the method by which an animal nose pattern image is analyzed to be used for identification, the order of which may be modified to better suit the equipment or circumstances: S 4001 acquisition of the subject animal's nose pattern image by utilizing the body stabilizer unit and image acquisition unit; S 4003 setting the ROI on the (processed) nose pattern image; S 4005 generating a nose pattern code from the fixed ROI; S 4006 enrolling the generated nose pattern code; S 4007 comparing the stored nose pattern code from the enrollment to the newly generated nose pattern code in one-to-one matching for verification; and S 4008 running one-to-many matching for identification. Images acquired from S 4001 that have been processed are called processed nose pattern images, and an additional step S 4002 for storing them may be included. Also, the step S 4004 that generates a standardized ROI from the ROI selected in S 4003 may also need to occur.
  • the image processing unit processes the acquired nose pattern images in order to increase the identification rate, and stores the resulting image.
  • Raw acquired images may present different levels of noise and blurring, and may require contrast adjustments to normalize the distribution of pixel values.
  • the present invention uses the histogram equalization technique to normalize the distribution of pixel values of images. In order to adjust the distribution of pixel values, a distribution function is fixed and histogram equalization is applied to each nose pattern image to have the same fixed distribution function.
  • Image filtering techniques may also be applied to take care of the noise and blurring issues, with Gaussian or median filters for noise level adjustment, and with a variety of low-pass filters in the frequency domain.
  • sharpening techniques using derivatives can be used to accentuate the embossed nose patterns
  • de-convolution techniques can be used to restore damaged images.
  • nose pattern images except when necessary to distinguish between the raw nose pattern images and the processed nose pattern images, both will be commonly referred to as nose pattern images for the sake of simplicity.
  • FIG. 41 is a schematic diagram briefly illustrating the ROI fixing unit as one embodiment of the present invention.
  • the ROI fixing unit may comprise the segmentation unit 4101 , the curve approximation unit 4102 , and the ROI dividing unit 4103 .
  • the segmentation unit sets the boundaries of the nostrils, which become the basis for setting the ROI in the nose pattern image.
  • FIG. 42 is a schematic diagram illustrating how to find the nostril boundary as one embodiment of the present invention.
  • FIG. 42 illustrates the nostril boundary setting process, where the nostrils appear as a shade due to the indirect illumination.
  • the boundary of this shade is the basis for the nostril boundary, which may take the form of a circular or elliptical arc, etc.
  • the boundary points are located based on the change in brightness along the ray from the fixed center points. Points along the rays extending in various directions that display a sharp change in brightness are marked as candidate points, and the correct boundary points are found among those candidate points based on the statistical analysis of nostril shape and location.
  • the curve approximation unit approximates the boundary curves of nostril boundaries using the boundary points found in the segmentation unit.
  • the final approximation curve is the best curve fitting the boundary points found by various regression analyses, and it is usually a circular arc or elliptical arc.
  • FIG. 43 is a diagram illustrating how to approximate the nostril boundaries with circles or ellipses as one embodiment of the present invention.
  • the left and right nostril boundaries can be regarded as symmetric curves when they are seen from the front of the nose, the two approximation curves can be asymmetric ellipses if the nose pattern image is taken from askew.
  • the two approximation curves can have different shapes resulting in that one curve is a circle, and the other an ellipse. It is also possible that the two approximation curves are different in size although they are all either circles or ellipses.
  • the ROI dividing unit extracts a quadrilateral region of a nose pattern image between the two approximation curves obtained from the approximation unit. This process consists of two steps: a) the region between two approximation curves is identified and b) a quadrilateral region contained in the identified region is extracted.
  • FIG. 44 is a schematic diagram illustrating how to identify the region between the two approximation curves (circles or ellipses) as one embodiment of the present invention.
  • two points which are on the intersections between each approximation curve and the line segment connecting two centers of the approximation curves are located, and the two tangent lines which tangent at each located point to the approximation curve (the left tangent line is denoted by T_L, and the right tangent line by T_R) are found.
  • These tangent lines may be perpendicular to the line segment connecting the two centers when the two approximation curves are symmetrical, and may not be perpendicular when they are not symmetrical.
  • the two connecting lines are then found: one line connecting two upper vertex points of the approximation curves and the other line connecting two lower vertex points (the upper line is denoted by T_U, and the lower line denoted by T_D).
  • the two connecting lines are tangent lines which tangent to the both of the approximation curves when they are both circles and the two lines connect two upper vertex points or two lower vertex points when they are both ellipses.
  • FIG. 45 is a schematic diagram illustrating how to extract the quadrilateral region between the two approximation curves as one embodiment of the present invention.
  • the ROI is the quadrilateral region encompassed by four lines obtained in Step A.
  • the shape and the size of the ROI may be varied depending on the relative position of the nose to the position of the image acquisition unit when the nose image is captured, and thus even the ROI from the same subject animal may be varied.
  • the two approximation curves may be obtained so that the line segment connecting the center points of the approximation curves passes the vertex points of the two approximation curves when they are both approximated by ellipses.
  • the line segment connecting the two center points of the two elliptical nostril boundary curves should pass the vertex point of each ellipse.
  • the boundary curves can be approximated by ellipses so that the line segment connecting the center points of the ellipses passes the vertex points of the ellipses.
  • FIG. 46 is a diagram illustrating how the ROI from the same nose pattern image may be varied depending on the approximation curves of the nostril boundaries. As shown in FIG. 46 , the quadrilateral ROI from even the same nose pattern image may be varied when different approximation curves are used, and the above quadrilateral ROI from even the same subject animal may also be varied depending on the relative position of the nose to the image acquisition unit during capture.
  • the standardized ROI fixing unit takes care of the transformation process of the ROI into the standard rectangular shape based on Equation (2).
  • FIG. 47 is a diagram illustrating the transformation process of the previously determined ROI into a standardized ROI as one embodiment of the present invention. As shown in FIG. 47 , a quadrilateral ROI with four vertices O, A, B, C is transformed into a rectangular area of width W and height H by Equation 2.
  • the coordinates of the corresponding point X in the ROI may not be integral values in contrast to the point in the standardized ROI which has integral coordinates a and b.
  • ROIs In the present invention, except when necessary to distinguish between the ROI and the standardized ROI, both will be commonly referred to as ROIs for the sake of simplicity.
  • ROIs A detailed account of the nose pattern code generation unit is given below.
  • FIG. 48 is a simplified block diagram describing the nose pattern code generation unit as one embodiment of the present invention.
  • the nose pattern code generation unit may comprise the frequency transform code generation unit 4801 and the masking code generation unit 4802 .
  • a nose pattern code consists of a frequency transform code and a masking code which are generated by the frequency transform code generation unit and the masking code generation unit, respectively.
  • a detailed account of the nose pattern code generation method is given below.
  • FIG. 49 is a block diagram illustrating the process of generating nose pattern codes from the ROI.
  • a nose pattern code consists of the frequency transform code generated in the frequency transform code generation unit and the masking code generated in the masking code generation unit using the ROI.
  • the nose pattern code is a 2-bit array and its component value is determined by predetermined frequency transform methods and parameters of the transforms.
  • the predetermined frequency transform methods may include several frequency methods including Gabor transform, Haar transform, Gabor Cosine transform, Gabor Sine transform, Sine transform, Cosine transform, and various wavelet transforms.
  • different frequencies for real and imaginary parts of Gabor transform may be used.
  • either of the real part of Gabor transform (Gabor Cosine transform) or the imaginary part of Gabor transform (Gabor Sine transform) may be used alone.
  • the choice of frequency transform methods in the nose pattern code generation unit may be determined according to the performance and the processing speed of the image recognition unit.
  • FIG. 50 is a diagram illustrating how to divide the ROI into smaller regions with specified dimensions which are called cell blocks and how frequency transform codes are generated from those cell blocks.
  • Each cell block may consist of one or more pixels.
  • a group of pixels may be reduced into one pixel by averaging the values in the group.
  • the group of pixels may be regarded as a cell block in this process.
  • each cell block may be represented by one pixel value using proper methods.
  • the total number of cell blocks and the size of each cell block may be varied depending on the size of the nose image, the breed of the subject animal, the frequency transform methods, parameters used in the frequency transform methods, etc.
  • the frequency transform codes consist of frequency transform values, each of which is obtained from a group of cell blocks called a cell-group as shown in FIG. 50 .
  • a cell-group is the basic unit for obtaining frequency transform codes.
  • Two different cell-groups may include some common cell blocks.
  • Each frequency transform value from a cell-group is a binary bit value (0 or 1) calculated based on the predetermined frequency transform method and parameters.
  • each cell-group gives only one frequency transform value so that the length of the frequency transform code is equal to the number of cell-groups
  • multiple frequency transform values may be obtained from one cell-group with multiple frequency transform methods and parameters. Also, with some frequency transform method, multiple frequency transform values may be obtained from each cell-group even with one parameter.
  • Equation (4) The frequency transformation method of Gabor transform is given by Equation (4) with its parameters.
  • Gabor Cosine transform and Gabor Sine transform each calculates the binary bit values of frequency transform codes using Equation (5) and Equation (6), respectively.
  • V ⁇ 0 W ⁇ ⁇ 0 H ⁇ I ⁇ ( a , b ) ⁇ e - i ⁇ ⁇ ⁇ x ⁇ ( a - a 0 ) ⁇ e - i ⁇ ⁇ ⁇ ⁇ y ⁇ ( b - b 0 ) ⁇ e - ( a - a 0 ) 2 ⁇ 2 ⁇ ⁇ e - ( b - b 0 ) 2 ⁇ 2 ⁇ d ⁇ ⁇ b ⁇ d ⁇ ⁇ a ( Equation ⁇ ⁇ 4 )
  • Re ⁇ ( V ) ⁇ 0 W ⁇ ⁇ 0 H ⁇ I ⁇ ( a , b ) ⁇ e - ( a - a 0 ) 2 ⁇ 2 ⁇ e - ( b - b 0 ) 2 ⁇ 2 ⁇ cos ⁇ ( ⁇ x ⁇ ( a
  • I(a,b) denotes the brightness of the pixel in the ROI at the position of (a,b) and represent coordinates of a point in the ROI.
  • ⁇ , ⁇ are parameters of Gabor transform to determine how large is the effective region to consider when calculating the frequency transform, and are parameters to determine the horizontal frequency and the vertical frequency, respectively.
  • FIG. 51 is a diagram illustrating how the two regions differ when using Gabor transform, Gabor Cosine transform, Gabor Sine transform as one embodiment of the present invention.
  • the theoretical region of integration in Equations (4), (5) and (6) is the whole ROI, but the actual region to consider when approximating the integration is restricted to the region where the value of
  • cell-groups may be formed so that each cell-group only consists of cell blocks where the value of
  • Such cell blocks are determined by the point (a 0 , b 0 ) and parameters ⁇ , ⁇ .
  • the region of such a cell block is denoted as R(a 0 , b 0 , ⁇ , ⁇ ) below.
  • R(a 0 , b 0 , ⁇ , ⁇ ) the region of such a cell block is denoted as R(a 0 , b 0 , ⁇ , ⁇ ) below.
  • the actual region of integration in Equations (4), (5) and (6) is the region R(a 0 , b 0 , ⁇ , ⁇ ).
  • the rectangle with dotted lines represents the cell-group, R(a 0 , b 0 , ⁇ , ⁇ ).
  • Equations (4), (5) and (6) can be approximated by Equations (4-1), (5-1) and (6-1).
  • V ⁇ ⁇ R ⁇ ( a 0 , b 0 , ⁇ , ⁇ ) ⁇ I ⁇ ( a , b ) ⁇ e - ⁇ i ⁇ ⁇ ⁇ x ⁇ ( a - a 0 ) ⁇ e - ⁇ i ⁇ ⁇ ⁇ ⁇ y ⁇ ( b - b 0 ) ⁇ e - ( a - a 0 ) 2 ⁇ 2 ⁇ e - ( b - b 0 ) 2 ⁇ 2 ⁇ d ⁇ ⁇ bd ⁇ ⁇ a ( Equation ⁇ ⁇ 4 ⁇ - ⁇ 1 )
  • Re ⁇ ( V ) ⁇ ⁇ R ⁇ ( a 0 , b 0 , ⁇ , ⁇ ) ⁇ I ⁇ ( a , b ) ⁇ e - ( a - a 0 ) 2 ⁇
  • V in Equation (4-1) is evaluated on the region R(a 0 , b 0 , ⁇ , ⁇ ).
  • the binary bit value of the frequency transform code is determined by the sign of Re(V) in Equation (5-1): it is 1 if Re(V) ⁇ 0, and 0 otherwise.
  • the binary bit value of the frequency transform code is determined by the sign of Im(V) in Equation (6-1) in the same way.
  • N denotes the length of the frequency transform code. Since multiple frequency transform values can be obtained from each cell-group using different frequency methods and their parameters, the length of the frequency transform code, N, may be different from the number of cell-groups.
  • Each binary value of the frequency transform codes can be obtained based on its own predetermined frequency transform method and parameters. In other words, different frequency methods or different parameters may be used for each binary value of the frequency transform code. In this way, the various features of a nose pattern image may be better encoded into the frequency transform code, and therefore increase the accuracy rate of recognition in comparison with the case where every binary value is calculated by using the same frequency method or parameters.
  • This strategy can also be applied to each breed type. Although the same frequency transform method or parameters may be chosen regardless of breed type, the accuracy rate of recognition may be increased if different frequency methods or parameters are properly chosen for each breed type.
  • the best frequency transform method and parameters may be chosen when generating nose pattern codes based on the estimated size of nose pattern features in the given nose pattern image.
  • the best frequency transform method and parameters e.g., frequencies
  • a fine distinction between breeds may be achieved where the sizes of nose pattern features are significantly different.
  • Each bit value of a masking code corresponds to a bit value of a frequency transform code.
  • a frequency code of N bits is generated from N configurations of frequency transform methods and values of a 0 , b 0 , ⁇ , ⁇ , ⁇ x , ⁇ y , each bit value of a masking code is also computed from each of the N configurations.
  • the length of making codes is the same as the length of frequency transform codes.
  • the masking code generation process goes through a light-reflection masking step and an additional masking step. Depending on the methods of masking code generation, both steps or only one step may be applied.
  • the nose pattern may appear damaged in the acquired nose image. Such regions are marked as damaged regions in the masking code.
  • each value of the making code is assigned so that it can be distinguished from the value whether the corresponding frequency transform value is damaged or not due to the light reflections. For example, when the actual integration region R(a 0 , b 0 , ⁇ , ⁇ ) includes damaged regions due to light reflections as in FIG. 50 , the value of 0 is assigned to the corresponding masking bit value to mark the frequency transform value as a damaged one. Otherwise, the value of 1 is assigned to the corresponding masking bit value.
  • the region containing nose patterns in the nose images of subject animals can be damaged due to nose or facial hairs, long whiskers, or foreign substances attached to wet noses.
  • the value of 0 is assigned to the corresponding masking bit value to mark the frequency transform value as a damaged one. Otherwise, the value of 1 is assigned to the corresponding masking bit value.
  • the nose pattern code consisting of the above-mentioned frequency transform code and masking code is generated, it is stored in the nose pattern code DB in the image recognition unit.
  • FIG. 52 is a simplified block diagram describing the nose pattern code matching unit.
  • the nose pattern code matching unit may include the nose pattern code verification unit 5201 and the nose pattern code identification unit 5202 .
  • the nose pattern code verification unit performs verification (one-to-one matching) by comparing the nose pattern code generated for verification and the nose pattern code stored in the nose pattern code DB in the image recognition unit. Verification of the generated nose pattern code is performed by computing the dissimilarity (the distance) of two nose pattern codes using one of following matching methods: a) simple matching, b) shift matching and c) block-wise shift matching.
  • FIG. 53 is a diagram illustrating the method of nose pattern code verification through simple matching.
  • the whole ROI A is compared to the whole ROI B in the simple matching.
  • the distance between the two nose pattern codes corresponding to cell-groups from A 1 to An of ROI A and cell-groups from B 1 to Bn of ROI B is less than the given threshold, it is concluded that the two nose pattern images are taken from the same subject. Otherwise, it is concluded that the two images are taken from different subject animals.
  • FIG. 54 is a diagram illustrating a situation in which different ROIs have been selected from the same nose pattern image of an individual for matching. As shown in FIG. 54 , selecting the same ROI is critical in increasing the accuracy rate of recognition since the different ROIs from the same subject animal result in high distance, and thus in a false non-match in the simple matching. There is a high probability of error in simple matching if it is difficult to locate the same ROI from the nose pattern images.
  • the probability of getting a false non-match is high when ROI A and ROI B are compared in simple matching even though they are from the same subject animal. More specifically, when cell-group A 1 from ROI A and cell-group B 1 from ROI B are compared in simple matching, it is concluded that they are taken from different individuals.
  • FIG. 55 is a diagram illustrating the method of nose pattern code identification through shift matching as one embodiment of the present invention. As shown in FIG. 55 , the local region a in ROI A is compared to the local region b in ROI B. The nose pattern code generated from cell-groups through of the local region “a” is compared to the nose pattern code generated from cell-groups through of the local region “b”.
  • the distance between two nose pattern codes is computed for each pair of local region “a” in ROI A and local region “b” in ROI B.
  • the local regions a and b in each ROI By translating the local regions a and b in each ROI, multiple values of distance are computed. If the minimum of the multiple values of distance is less than the given threshold, it is concluded that the two nose pattern images are taken from the same individual. Otherwise, it is concluded that the two nose pattern images are taken from different subject animals.
  • nose pattern codes from all possible local regions in the ROI should be generated.
  • the value of frequency transform should be computed for each cell-group in all possible local regions.
  • it is required to compute all values of the frequency transform for all cell-groups in all possible local regions.
  • the values of frequency transform from pre-computed cell-groups may be used rather than computing every value of frequency transform for every cell-group.
  • the pre-computed value from one local region is used for the other local region rather than re-computing the value of frequency transform for the cell-group.
  • local regions and cell-groups may be constructed with the efficiency of this computation in mind.
  • all values of frequency transform from all possible cell-groups are computed first and the nose pattern code for each local region is constructed using the pre-computed values of frequency transform.
  • FIG. 56 is a diagram illustrating a matching situation in which the ROIs of the nose pattern code selected from the same individual have nonidentical vertical and horizontal proportions.
  • a slice region a partial region of a local region is selected and compared by translating the slice region in a local region. For example, as in FIG. 56 , if slice regions A 3 , A 5 , A 7 in local region “a” have the same nose pattern as slice regions B 2 , BS, B 8 in local region “b,” respectively, the corresponding values of frequency transform from those slice regions are the same. Thus, by translating a slice region, it is probable that a match will be made between ROI A and ROI B.
  • FIG. 57 is a diagram illustrating the method of nose pattern code identification through block-wise shift matching as one embodiment of the present invention.
  • slice region ⁇ of local region “a” in ROI A is compared to slice region ⁇ of local region “b” in ROI B.
  • Local region “a” in ROI A and local region “b” in ROI B are subdivided into n*m equal sized pieces with n horizontal pieces and m vertical pieces. Then the distance between the nose pattern code corresponding to each piece of local region “a” and local region “b” is computed.
  • slice regions a and 0 are translated in a given range so that multiple values of distance are computed.
  • the distance is computed for each pair of slice regions and thus multiple values of distance for each pair of local regions. So, one representative value of distance (called “final distance”) from multiple values of distance needs to be computed to compare with the given threshold.
  • the distance should be computed for each pair of slice regions.
  • the distance (called block distance) between a pair of slice regions is the minimum value of all possible distances computed from all possible translations of slice regions.
  • the final distance may be defined as one of the minimum, a geometric average, an arithmetic average of all block distances.
  • the above ROI A can be the ROI from which a stored nose pattern code is generated, and the ROI B can be the one from which the nose pattern code for verification is generated, and vice versa.
  • the nose pattern code for ROI A consists of N bits of frequency transform code C 1 and N bits of masking code M 1
  • the nose pattern code for ROI B consists of N bits of frequency transform code C 2 and N bits of masking code M 2 .
  • the distance (D) between two nose pattern codes can be computed by Equation (7).
  • Equation (7) XOR denotes the operator of bitwise Exclusive-OR, AND the operator of bitwise AND.
  • The number of bits whose value is 1 in the array of bits A is denoted by
  • the nose pattern code for local region “a” in ROI A consists of N bits of frequency transform code C 1 and N bits of masking code M 1
  • the nose pattern code for local region “b” in ROI B consists of N bits of frequency transform code C 2 and N bits of masking code M 2 .
  • FIG. 58 is a diagram illustrating the process of nose pattern code identification through shift matching using Gabor Sine transform.
  • R 1 be the set of all the points ( ⁇ x , ⁇ y ) of the lower left vertices of all the possible translated local regions in ROI A
  • R 2 the set of all the points ( ⁇ x , ⁇ y ) of the lower left vertices of all the possible translated local regions in ROI B.
  • one nose pattern code (a(1,1)) from ROI A and nine nose pattern codes (b(0,0), b(1,0), b(2,0), b(0,1), b(1,1), b(2,1), b(0,2), b(1,2), b(2,2)) from ROI B are generated, and these codes give a total of nine values of distance by one-by-one comparison. Thus, in this case, it is only necessary to see if the minimum of nine values of distance is less than the given threshold.
  • Equation (8) XOR denotes the operator of bitwise Exclusive-OR, AND the operator of bitwise AND.
  • The number of bits whose value is 1 in the array of bits A is denoted by
  • the nose pattern code for slice region ⁇ of local region “a” in ROI A consists of N bits of frequency transform code C 1 and N bits of masking code M 1
  • the nose pattern code for slice region ⁇ of local region “b” in ROI B consists of N bits of frequency transform code C 2 and N bits of masking code M 2 .
  • FIG. 59 is a diagram illustrating the process of nose pattern code identification through block-wise shift matching using Gabor Sine transform.
  • one nose pattern code ( ⁇ (1,1)) from slice region ⁇ in ROI A and nine nose pattern codes ( ⁇ (0,0), ⁇ (1,0), ⁇ (2,0), ⁇ (0,1), ⁇ (1,1), ⁇ (2,1), ⁇ (0,2), ⁇ (1,2), ⁇ (2,2)) from slice region ⁇ in ROI B are generated, and these codes give a total of nine values of distance by one-by-one comparison.
  • the minimum of all these values of distance is called the distance between slice regions, and the final distance between two ROIs may be defined as one of the arithmetic average, the geometric average, and the minimum of all possible distances between two corresponding slice regions.
  • D ⁇ ( k ) min ( ⁇ x , ⁇ y ) ⁇ R 1 , ( ⁇ x , ⁇ y ) ⁇ R 2 ⁇
  • Equation (9) XOR denotes the operator of bitwise Exclusive-OR, AND the operator of bitwise AND.
  • The number of bits whose value is 1 in the array of bits A is denoted by
  • the final distance between two ROIs may be defined as one of the geometric average, the arithmetic average, and the minimum of all values of distance between two corresponding slice regions for all pairs of slice regions.
  • the above ROI A can be the ROI from which a stored nose pattern code is generated, and the ROI B can be the one from which the nose pattern code for verification is generated, and vice versa.
  • the nose pattern code identification unit performs identification (one-to-many matching) by comparing the nose pattern code generated for identification with multiple nose pattern codes stored in the nose pattern code DB in the image recognition unit.
  • While verification requires computing the distance between a single nose pattern code and a single stored nose pattern code through simple matching, shift matching or block-wise shift matching, identification (one-to-many matching) requires computing the distances between a single nose pattern code and each of the multiple nose pattern codes stored in the DB.
  • FIG. 60 is a diagram illustrating the method of nose pattern code identification (one-to-many matching) as one embodiment of the present invention.
  • identification is performed by computing the distances between a single nose pattern code given for identification and each of k nose pattern codes (nose pattern code_ 1 , . . . , nose pattern code_k) stored in the DB.
  • each nose pattern code stored in the DB consists of the frequency transform code C and the masking code M, and the DB of pre-registered nose pattern codes may be constructed as a part of the image recognition unit.
  • a total of k distances are computed, and these k distances may be denoted by D 1 , . . . , Dk.
  • the values of D 1 , . . . , Dk may be sorted in a descending or ascending order.
  • One or more candidate nose pattern codes can be selected, according to one of three predefined selection rules: (a) a rule selecting the nose pattern code that gives the minimum distance among all those that yield distances less than the given threshold, (b) a rule selecting all the nose pattern codes whose distance is less than the given threshold, and (c) a rule selecting a predefined number, say n, of nose pattern codes whose distance is within top n least distances.
  • the image recognition unit according to an embodiment of the present invention as described above may be implemented as a program and recorded on a medium readable by a computer, which includes all kinds of compatible data storing devices such as ROM, RAM. CD ROM, magnetic tapes, floppy disks, and optical data storage devices, and carrier wave (e.g., transmission through the Internet).
  • the recording medium may also be distributed over network coupled computer systems so that the computer readable code can be stored and executed in a distributed manner.
  • the functional programs, codes, and code segments for implementing the present invention may be easily construed by programmers skilled in the area relevant to the invention.
  • the flowchart in FIG. 11 summarizes the method of animal recognition in the present invention, starting with S 1101 selecting and S 1102 fitting the animal into the appropriate body stabilizer unit; S 1103 fixing the nose of the subject onto the image acquisition unit and S 1104 acquiring the nose pattern image; S 1105 at the image recognition unit, generating a nose pattern code using the raw or processed nose pattern image and S 1106 enrolling and identifying the individual using the nose pattern code.
  • S 2101 select the appropriate body stabilizer unit for the subject animal by taking into consideration the overall size, leg length, feet size, head size, and the relative location of the nose;
  • S 2102 fit the subject animal into the upper body stabilizer unit;
  • S 2103 fasten the upper body by utilizing the upper body stabilizing brace subunit and upper body stabilizing brace pressure adjuster subunit to fit the shoulder width;
  • S 2104 fit the subject animal into the lower body stabilizer;
  • S 2105 fasten the lower body by utilizing the lower body stabilizing brace subunit and lower body stabilizing brace pressure adjuster subunit to fit the ankles or legs;
  • S 2106 set the stance width adjuster, and also the height adjuster to fit the subject's height if necessary to connect the upper and lower body stabilizer units;
  • S 2107 fasten the head by utilizing the head stabilizer unit, making sure to set the height adjuster unit to the correct height and the horizontal balance adjuster unit to have the nose facing the image acquisition unit head-on.
  • FIG. 38 An example of the image acquisition process is illustrated in FIG. 38 ; the order of events need not be limited as follows.
  • S 3801 the operator selects the species on the display unit to start the automatic mode, or chooses the manual mode;
  • S 3802 in automatic mode once the species selection is made, pressing the capture button starts the acquisition of the batch of nose pattern images at n frames per second while the lens module is shifted about within the preset range of positions (adjusting values of a and b).
  • the image capture unit, illumination unit and front unit are automatically adjusted to accommodate the subject animal based on the values stored in the parameter DB for FOV, focus, luminosity, etc.
  • manual mode the operator visually evaluates the features and variables of the nose pattern images through the display unit, and selects the best image.
  • FIG. 40 illustrates the method by which an animal nose pattern image is analyzed to be used for identification, the order of which may be modified to better suit the equipment or circumstances: S 4001 acquisition of the subject animal's nose pattern image by utilizing the body stabilizer unit and image acquisition unit; S 4003 setting the ROI on the (processed) nose pattern image; S 4005 generating a nose pattern code from the fixed ROI; S 4006 enrolling the generated nose pattern code; S 4007 comparing the stored nose pattern code from the enrollment to the newly generated nose pattern code in one-to-one matching for verification; and S 4008 running one-to-many matching for identification. Images acquired from S 4001 that have been processed are called processed nose pattern images, and an additional step S 4002 for storing them may be included. Also, the step S 4004 that generates a standardized ROI from the ROI selected in S 4003 may also need to occur.
  • the main processor unit selects nose pattern images that are of sufficient quality to be used in the image recognition unit, out of all the images captured by the image capture unit.
  • each image is given individual scores on specific variables, and images that pass the threshold set by the image analysis unit are selected. If none out of a particular batch meet the threshold, then that whole group is discarded and a request for a new batch is sent to the image capture unit.
  • the images are evaluated on such criteria as, the amount of light reflection, sharpness, contrast ratio, ROI for capture, noise level, etc; and only those images that pass the threshold for each variable are accepted.
  • the threshold the one with the highest total score (sum of individual scores) is selected, and this process may take place simultaneously as the image acquisition in the image analysis unit or the image recognition unit.
  • variables there are two types of variables: those that are not related to species-specific characteristics (A 1 -A 3 ) and those that are (A 4 -A 12 ).
  • the former includes sharpness A 1 , contrast A 2 , and noise level A 3 ;
  • the latter includes ROI for capture A 4 , presence of light reflection A 5 , nostril location A 6 , sharpness of nostril image A 7 , contrast level of nostril image A 5 , noise level of nostril image A 9 , sharpness of the border between the nostril and ROI A 10 , contrast level at the border between the nostril and ROI A 11 , and noise level at the border between the nostril and ROI A 12 .
  • Variables may be appropriately added to or subtracted from the above list depending on a subject animal species' particular characteristics.
  • each image in a batch is given individual scores on specific variables, and images that pass the threshold set by the image analysis unit are selected.
  • images that pass the threshold set by the image analysis unit are selected.
  • the threshold the one with the highest total score (sum of individual scores) is selected.
  • Freshly acquired nose pattern images may require processing in order to increase the identification rate.
  • Raw acquired images may present different levels of noise and blurring, and may require contrast adjustments to normalize the distribution of pixel values.
  • the present invention uses the histogram equalization technique to normalize the distribution of pixel values of images.
  • a distribution function is fixed and applied with histogram equalization for each nose pattern image to have the same fixed distribution function.
  • Image filtering techniques may also be applied to take care of the noise and blurring issues, with Gaussian or median filters for noise level adjustment, and a variety of low-pass filters in the frequency domain. Moreover, sharpening techniques using derivatives can be used to accentuate the embossed nose patterns, and de-convolution techniques can be used to restore damaged images.
  • the nostril boundary needs to be found in order to fix the ROI from the nose pattern image.
  • FIG. 42 is a schematic diagram illustrating how to find the nostril boundary as one embodiment of the present invention.
  • FIG. 42 illustrates the nostril boundary setting process, where the nostrils appear as a shade due to the indirect illumination.
  • the boundary of this shade is the basis for the nostril boundary, which may take the form of a circular or elliptical arc, etc.
  • the boundary points are located based on the change in brightness along the ray from the fixed center points. Points along the rays extending in various directions that display a sharp change in brightness are marked as candidate points, and the correct boundary points are found among those candidate points based on the statistical analysis of nostril shape and location.
  • the nostril boundaries are approximated by curves fitting the boundary points found in the above step.
  • the final approximation curve is the best curve fitting the boundary points found by various regression analyses, and it is usually a circular arc or elliptical arc.
  • the left and right nostril boundaries can be regarded as symmetric curves when they are seen from the front of the nose, the two approximation curves can be asymmetric ellipses if the nose pattern image is taken from askew.
  • the two approximation curves can have different shapes resulting in that one curve is a circle, and the other an ellipse. It is also possible that the two approximation curves are different in size although they are all either circle or ellipses.
  • a quadrilateral region of a nose pattern image between the two approximation curves obtained from the above step is to be defined.
  • This process consists of two steps: a) the region between two approximation curves is identified and b) a quadrilateral region contained in the identified region is extracted.
  • FIG. 44 is a schematic diagram illustrating how to identify the region between the two approximation curves (circles or ellipses) as one embodiment of the present invention.
  • two points which are on the intersections between each approximation curve and the line segment connecting two centers of the approximation curves are located, and the two tangent lines which tangent at each located point to the approximation curve (the left tangent line is denoted by T_L, and the right tangent line by T_R) are found.
  • These tangent lines may be perpendicular to the line segment connecting the two centers when the two approximation curves are symmetrical, and may not be perpendicular when they are not symmetrical.
  • the two connecting lines are then found: one line connecting two upper vertex points of the approximation curves and the other line connecting two lower vertex points (the upper line is denoted by T_U, and the lower line denoted by T_D).
  • the two connecting lines are tangent lines which tangent to the both of the approximation curves when they are both circles and the two lines connects two upper vertex points or two lower vertex points when they are both ellipses.
  • FIG. 45 is a schematic diagram illustrating how to extract the quadrilateral region between the two approximation curves as one embodiment of the present invention.
  • the ROI is the quadrilateral region encompassed by four lines obtained in Step A.
  • the shape and the size of the ROI may be varied depending on the relative position of the nose to the position of the image acquisition unit when the nose image is captured, and thus even the ROI from the same animal subject may be varied.
  • the two approximation curves may be obtained so that the line segment connecting two center points of the approximation curves passes each of vertex points of two approximation curves when they are both approximated by ellipses.
  • the line segment connecting the two center points of the two elliptical nostril boundary curves should pass the vertex point of each ellipse.
  • the boundary curves can be approximated by ellipses so that the line segment connecting two center points of ellipses passes the vertex points of ellipses.
  • FIG. 46 is a diagram illustrating how the ROI from the same nose pattern image may be varied depending on the approximation curves of the nostril boundaries. As shown in FIG. 46 , the quadrilateral ROI from even the same nose pattern image may be varied when different approximation curves are used, and the above quadrilateral ROI from even the same subject animal may also be varied depending on the relative position of the nose to the image acquisition unit during capture.
  • the standardized ROI fixing unit takes care of the transformation process of the ROI into the standard rectangular shape based on Equation (2).
  • Nose pattern codes are generated via steps of generating the a) frequency transform code and b) masking code from the ROI.
  • FIG. 49 is a block diagram illustrating the process of generating nose pattern codes from the ROI.
  • a nose pattern code consists of the frequency transform code and the masking code generated from the whole ROI.
  • the nose pattern code is a 2-bit array and its component value is determined by predetermined frequency transform methods and parameters of the transforms.
  • the predetermined frequency transform methods may include several frequency methods including Gabor transform, Haar transform, Gabor Cosine transform, Gabor Sine transform, Sine transform, Cosine transform, and various wavelet transforms.
  • different frequencies for real and imaginary parts of Gabor transform may be used.
  • either of the real part of Gabor transform (Gabor Cosine transform) or the imaginary part of Gabor transform (Gabor Sine transform) may be used alone.
  • the choice of frequency transform methods may be determined based on the performance and the processing speed of the image recognition unit. The technical details regarding the generation of frequency transform codes have been omitted to avoid redundancy.
  • Each bit value of a masking code corresponds to a bit value of a frequency transform code.
  • each bit value of a masking code is also computed from each of the N configurations.
  • the length of making codes is the same as the length of frequency transform codes.
  • the masking code generation process goes through a light-reflection masking step and an additional masking step. Depending on the methods of masking code generation, both steps or only one step may be applied. The technical details regarding the generation of masking codes have been omitted to avoid redundancy.
  • the nose pattern code verification unit performs verification (one-to-one matching) by comparing the nose pattern code generated for verification and the stored nose pattern code. Verification of the generated nose pattern code is performed by computing the distance between two nose pattern codes through one of the following matching methods: a) simple matching, b) shift matching and c) block-wise shift matching. The technical details regarding the matching methods have been omitted to avoid redundancy.
  • While verification requires computing the distance between a single nose pattern code and a single stored nose pattern code through simple matching, shift matching or block-wise shift matching, identification (one-to-many matching) requires computing the distances between a single nose pattern code and each of the multiple nose pattern codes stored in the DB.
  • FIG. 60 is a diagram illustrating the method of nose pattern code identification (one-to-many matching) as one embodiment of the present invention.
  • identification is performed by computing the distances between a single nose pattern code given for identification and each of k nose pattern codes (nose pattern code_ 1 , . . . , nose pattern code_k) stored in the DB.
  • each nose pattern code stored in the DB consists of the frequency transform code C and the masking code M, and the DB of pre-registered nose pattern codes may be constructed as a part of the image recognition unit.
  • a total of k distances are computed, and these k distances may be denoted by D 1 , . . . , Dk.
  • the values of D 1 , . . . . Dk may be sorted in a descending or ascending order.
  • One or more candidate nose pattern codes can be selected, according to one of three predefined selection rules: (a) a rule selecting the nose pattern code that gives the minimum distance among all those that yield distances less than the given threshold, (b) a rule selecting all the nose pattern codes whose distance is less than the given threshold, and (c) a rule selecting a predefined number, say n, of nose pattern codes whose distance is within top n least distances.
  • the present invention relates to the apparatus and method of animal recognition using nose patterns. Specifically, it involves the use of a body stabilizer unit, image acquisition unit, and image recognition unit to obtain identifiable nose pattern images.
  • the proposed method is physically and economically easy to implement and requires little expertise on the part of the operator; and as such, presents a very high potential for industrial applicability.

Landscapes

  • Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • Medical Informatics (AREA)
  • Environmental Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Animal Behavior & Ethology (AREA)
  • Public Health (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Surgery (AREA)
  • Biomedical Technology (AREA)
  • Pathology (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Veterinary Medicine (AREA)
  • Multimedia (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Animal Husbandry (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • Human Computer Interaction (AREA)
  • Artificial Intelligence (AREA)
  • Zoology (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Birds (AREA)
  • Radiology & Medical Imaging (AREA)
  • Physical Education & Sports Medicine (AREA)
  • Otolaryngology (AREA)
  • Databases & Information Systems (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Quality & Reliability (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
US14/893,043 2013-05-22 2014-05-20 Device and method for recognizing animal's identity by using animal nose prints Active 2035-09-22 US10268880B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US16/185,354 US10552673B2 (en) 2013-05-22 2018-11-09 Device and method for recognizing animal's identity by using animal nose prints

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
KR10-2013-0057667 2013-05-22
KR1020130057667A KR101527801B1 (ko) 2013-05-22 2013-05-22 동물들의 코무늬를 이용한 동물 개체 인식 장치 및 방법
PCT/KR2014/004487 WO2014189250A2 (ko) 2013-05-22 2014-05-20 동물들의 코무늬를 이용한 동물 개체 인식 장치 및 방법

Related Parent Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2014/004487 A-371-Of-International WO2014189250A2 (ko) 2013-05-22 2014-05-20 동물들의 코무늬를 이용한 동물 개체 인식 장치 및 방법

Related Child Applications (2)

Application Number Title Priority Date Filing Date
US16/185,354 Division US10552673B2 (en) 2013-05-22 2018-11-09 Device and method for recognizing animal's identity by using animal nose prints
US16/185,324 Division US10565440B2 (en) 2013-05-22 2018-11-09 Device for recognizing animal's identity by using animal nose prints

Publications (2)

Publication Number Publication Date
US20160259970A1 US20160259970A1 (en) 2016-09-08
US10268880B2 true US10268880B2 (en) 2019-04-23

Family

ID=51934284

Family Applications (3)

Application Number Title Priority Date Filing Date
US14/893,043 Active 2035-09-22 US10268880B2 (en) 2013-05-22 2014-05-20 Device and method for recognizing animal's identity by using animal nose prints
US16/185,354 Active 2034-05-23 US10552673B2 (en) 2013-05-22 2018-11-09 Device and method for recognizing animal's identity by using animal nose prints
US16/185,324 Active 2034-06-13 US10565440B2 (en) 2013-05-22 2018-11-09 Device for recognizing animal's identity by using animal nose prints

Family Applications After (2)

Application Number Title Priority Date Filing Date
US16/185,354 Active 2034-05-23 US10552673B2 (en) 2013-05-22 2018-11-09 Device and method for recognizing animal's identity by using animal nose prints
US16/185,324 Active 2034-06-13 US10565440B2 (en) 2013-05-22 2018-11-09 Device for recognizing animal's identity by using animal nose prints

Country Status (7)

Country Link
US (3) US10268880B2 (de)
EP (2) EP3029603B1 (de)
JP (3) JP6466920B2 (de)
KR (1) KR101527801B1 (de)
CA (1) CA2925275C (de)
ES (1) ES2936063T3 (de)
WO (1) WO2014189250A2 (de)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200250733A1 (en) * 2019-02-01 2020-08-06 Societe Des Products Nestle S.A. Pet food recommendation devices and methods

Families Citing this family (47)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3002744B1 (de) * 2013-03-12 2018-10-24 Arizona Board of Regents, a Body Corporate of the State of Arizona acting for and on behalf of Arizona State University Fälschungssicherung von etiketten mittels bildverarbeitung von dendritischen strukturen als physikalisch nicht klonbare funktion
US10810731B2 (en) 2014-11-07 2020-10-20 Arizona Board Of Regents On Behalf Of Arizona State University Information coding in dendritic structures and tags
KR102375864B1 (ko) * 2015-02-10 2022-03-18 한화테크윈 주식회사 요약 영상 브라우징 시스템 및 방법
FI3316680T4 (fi) 2015-07-01 2023-11-02 Viking Genetics Fmba Järjestelmä ja menetelmä yksittäisten eläinten tunnistamiseksi selästä otettujen kuvien perusteella
US10621445B2 (en) * 2016-06-29 2020-04-14 Toyota Motor Engineering & Manufacturing North America, Inc. Vehicle video systems including rear mounted video cameras
CN109496123B (zh) * 2016-07-20 2022-02-08 农场机器人和自动化有限公司 家畜的机器人协助监视
US11430233B2 (en) 2017-06-16 2022-08-30 Arizona Board Of Regents On Behalf Of Arizona State University Polarized scanning of dendritic identifiers
KR102083509B1 (ko) * 2017-07-04 2020-03-04 심동희 동물병원 경영 클라우드 플랫폼의 운영 장치 및 그 기반으로 구축한 비문 기반 동물 개체 인식 방법
US11025803B2 (en) * 2017-10-11 2021-06-01 Iscilab Corporation Apparatus for capturing animal nose pattern images on mobile devices
KR102168508B1 (ko) * 2017-12-05 2020-10-21 (주)링크옵틱스 비문촬영시스템
US10755404B2 (en) * 2017-12-07 2020-08-25 International Business Machines Corporation Integrated circuit defect detection using pattern images
EP3494780A1 (de) * 2017-12-07 2019-06-12 Siemens Aktiengesellschaft Verfahren und anordnung zur tieridentifikation
JP7362234B2 (ja) * 2017-12-22 2023-10-17 東芝ライテック株式会社 支援システム
EP3732662B1 (de) * 2017-12-29 2023-10-11 InterDigital VC Holdings, Inc. Verfahren und system zur bewahrung der farbkalibrierung mittels gemeinsamer objekte
US10535006B2 (en) 2018-01-23 2020-01-14 Here Global B.V. Method, apparatus, and system for providing a redundant feature detection engine
CN108830138B (zh) * 2018-04-26 2021-05-07 平安科技（深圳）有限公司 牲畜识别方法、装置及存储介质
WO2019210129A1 (en) 2018-04-26 2019-10-31 Kozicki Michael N Fabrication of dendritic structures and tags
CN110826371A (zh) * 2018-08-10 2020-02-21 京东数字科技控股有限公司 一种动物识别方法、装置、介质及电子设备
KR20200019302A (ko) * 2018-08-13 2020-02-24 주식회사 아이싸이랩 동물의 비문 이미지 획득 방법
KR102170931B1 (ko) 2018-08-29 2020-10-29 주식회사 아이싸이랩 비문 획득을 위한 조명 장치
KR102131007B1 (ko) * 2018-09-11 2020-07-09 주식회사 아이싸이랩 동물 등록 및 인증 중 적어도 하나의 서비스 제공 방법
KR102117860B1 (ko) * 2018-10-11 2020-06-02 주식회사 핏펫 동물의 안면 및 비문 이미지에 기초하여 동물의 개체 정보를 제공하는 컴퓨터 프로그램 및 단말기
KR102228725B1 (ko) * 2018-10-17 2021-03-17 여태환 동물의 생체인식 식별을 위한 방법 및 장치
CN109858363B (zh) * 2018-12-28 2020-07-17 北京旷视科技有限公司 一种狗鼻纹特征点的检测方法、装置、系统及存储介质
CN109948458A (zh) * 2019-02-25 2019-06-28 广东智媒云图科技股份有限公司 基于鼻纹的宠物身份识别方法、装置、设备及存储介质
CN110532854B (zh) * 2019-07-11 2021-11-26 中国农业大学 一种生猪爬跨行为检测方法及系统
US20210049355A1 (en) * 2019-08-16 2021-02-18 Stephanie Sujin CHOI Method for clustering and identifying animals based on the shapes, relative positions and other features of body parts
US11051493B2 (en) * 2019-09-09 2021-07-06 Council Of Agriculture Method and system for distinguishing identities based on nose prints of animals
US11450130B2 (en) 2019-09-25 2022-09-20 Pal Universe, Inc. Animal identification based on unique nose patterns
CN110762031B (zh) * 2019-10-23 2020-08-14 台州辉腾泵业有限公司 基于数据定量控制的水泵驱动装置
CN110929650B (zh) * 2019-11-25 2022-04-12 蚂蚁胜信(上海)信息技术有限公司 豢养物身份识别方法、装置、计算设备、可读存储介质
KR102251434B1 (ko) * 2019-12-24 2021-05-12 이주철 애완동물 식별과 생애주기 관리 인프라시스템 및 활용방법
KR102215133B1 (ko) * 2019-12-24 2021-02-10 이주철 반려동물의 비문 촬영 장치
KR102318500B1 (ko) 2020-02-28 2021-10-29 주식회사 아이싸이랩 카메라가 내장된 단말기를 이용한 동물 코무늬 이미지의 등록, 인증 및 조회 서비스 방법
US20210315186A1 (en) * 2020-04-14 2021-10-14 The United States Of America, As Represented By Secretary Of Agriculture Intelligent dual sensory species-specific recognition trigger system
WO2021230680A1 (en) * 2020-05-13 2021-11-18 Pireco Co,. Ltd. Method and device for detecting object in image
CN111831843A (zh) * 2020-06-11 2020-10-27 广东工业大学 一种基于区块链的动物身份管理方法和系统
KR102176934B1 (ko) * 2020-09-24 2020-11-10 주식회사 디랩스 안면인식을 이용한 반려동물 종합 관리 서비스 제공 시스템
EP4229546A4 (de) 2020-10-14 2024-04-10 One Cup Productions Ltd. Visuelle identifizierungs-, verfolgungs-, überwachungs- und bewertungssysteme für tiere und verfahren dafür
KR20220087310A (ko) 2020-12-17 2022-06-24 대한민국(농촌진흥청장) 영상 기반 반려견 개체 식별 방법 및 장치
KR102303862B1 (ko) * 2020-12-28 2021-09-17 한국축산데이터 주식회사 농업회사법인 인공지능 기반 온라인 가축 경매 장치 및 방법
USD977509S1 (en) * 2021-03-11 2023-02-07 Mars, Incorporated Display screen or portion thereof with a graphical user interface
USD977521S1 (en) * 2021-03-11 2023-02-07 Mars, Incorporated Display screen or portion thereof with a graphical user interface
KR102325259B1 (ko) * 2021-05-11 2021-11-11 (주) 아지랑랑이랑 반려동물 생애 관리 시스템 및 그 방법
WO2023095174A1 (en) * 2021-11-28 2023-06-01 Adis Technologies Pvt. Ltd. A process for identification of a snouted animal
JP7464305B2 (ja) * 2021-11-29 2024-04-09 まちなかＭｅセンター株式会社 動物のバイタルサイン管理システム、及び、これを備えた動物のバイタルサイン管理方法
CN115797226B (zh) * 2023-01-09 2023-04-25 腾讯科技（深圳）有限公司 图像降噪方法、装置、计算机设备和存储介质

Citations (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5690365A (en) 1996-06-03 1997-11-25 Fascenelli, Jr.; Frederick W. Pet identification tags
JPH1145330A (ja) 1997-07-29 1999-02-16 Fujitsu Denso Ltd 鼻紋採取装置
US6003473A (en) 1998-02-12 1999-12-21 Printz; Robert L. Pet identification and retrieval device
JP2001148957A (ja) 1999-11-25 2001-06-05 Nec Yonezawa Ltd 鼻紋認証装置による家畜管理システム
USD458421S1 (en) 2001-08-23 2002-06-04 Dennis Jay Wasserman Pet identification tag
KR20020066771A (ko) 2001-02-13 2002-08-21 박현종 네트워크 기반의 개의 비문 정보를 이용한 신원확인 시스템
US6568354B1 (en) 2002-02-05 2003-05-27 Dennis J. Wasserman Pet identification tag with electronic display
JP2003346148A (ja) 2002-05-23 2003-12-05 Surge Miyawaki Co Ltd 鼻紋の照合方法
KR200337597Y1 (ko) 2003-10-20 2004-01-03 김주경 동물의 유동을 방지하는 유동방지기구
KR20040006822A (ko) 2002-07-15 2004-01-24 김윤옥 인터넷 망과 네트워크 망을 통한 소 개체관리 시스템 및제어방법
KR20040008072A (ko) 2002-07-15 2004-01-28 김윤옥 소 개체관리 휴대형 정보 단말기 및 비문인식방법
USD491323S1 (en) 2003-04-14 2004-06-08 Amy Wilcox Pet identification collar
US20040119831A1 (en) * 1999-12-03 2004-06-24 Yutaka Miyawaki Method of management of pictures and method of recording pictures of animals
USD494326S1 (en) 2002-12-23 2004-08-10 Angela Rae Long Personalized animal/pet identification tag
US6845382B2 (en) 2000-12-22 2005-01-18 Louis B. Meadows Pet identification system and method
JP2007135501A (ja) 2005-11-21 2007-06-07 Atom System:Kk 鼻特徴情報生成装置及び鼻特徴情報生成プログラム
US7230539B2 (en) 2004-10-26 2007-06-12 Klein Elliot S Pet identification tag with internal cavity for transponder capsule
US20150078626A1 (en) * 2013-09-17 2015-03-19 William Brian Kinard Animal / pet identification system and method based on biometrics
US20160110383A1 (en) * 2011-07-11 2016-04-21 Snoutscan Llc System and method for animal identification

Family Cites Families (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3250252A (en) * 1964-05-28 1966-05-10 Leopold Seymour Veterinary animal restraining apparatus
US4286547A (en) * 1980-01-28 1981-09-01 Nuwbauer Robert E Animal restraint unit
US4693212A (en) * 1986-02-03 1987-09-15 Black Kenneth R Animal restraint apparatus and method
US5799617A (en) * 1995-09-27 1998-09-01 Rena Pimpl Animal head support device and method of use
JP3563184B2 (ja) * 1995-11-22 2004-09-08 畜産用電子技術研究組合 家畜の鼻紋採取装置
JPH09265523A (ja) * 1996-03-28 1997-10-07 Fujitsu Denso Ltd 鼻紋採取装置及び鼻紋採取処理方法
US6538243B1 (en) * 2000-01-04 2003-03-25 Hewlett-Packard Company Contact image sensor with light guide having least reflectivity near a light source
JP2007512830A (ja) * 2003-12-03 2007-05-24 サン−ヒー パク ペット用美容、医療及び浴槽システム
JP4080476B2 (ja) * 2004-10-29 2008-04-23 岩則 中田 犬保持装置
WO2007077694A1 (ja) * 2005-12-28 2007-07-12 Atomsystem Co., Ltd. 鼻紋情報生成装置及び鼻紋情報照合装置及び鼻紋情報生成プログラム及び鼻紋情報生成方法
US20080298642A1 (en) * 2006-11-03 2008-12-04 Snowflake Technologies Corporation Method and apparatus for extraction and matching of biometric detail
JP4665997B2 (ja) * 2008-05-12 2011-04-06 日本電気株式会社 生体パターン撮像装置、生体パターン撮像方法、生体パターン撮像プログラム
US20110006172A1 (en) * 2009-07-09 2011-01-13 Sargent Timothy J Animal Head Positioning Device
JP5437016B2 (ja) * 2009-10-22 2014-03-12 株式会社ロゼフテクノロジー 画像処理用照明装置
JP5590412B2 (ja) * 2011-02-15 2014-09-17 株式会社デンソー 投光システム及び顔撮影システム
US9826713B2 (en) * 2015-09-28 2017-11-28 Hadi Hosseini Animal muzzle pattern scanning device

Patent Citations (23)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5690365A (en) 1996-06-03 1997-11-25 Fascenelli, Jr.; Frederick W. Pet identification tags
JPH1145330A (ja) 1997-07-29 1999-02-16 Fujitsu Denso Ltd 鼻紋採取装置
US6003473A (en) 1998-02-12 1999-12-21 Printz; Robert L. Pet identification and retrieval device
JP2001148957A (ja) 1999-11-25 2001-06-05 Nec Yonezawa Ltd 鼻紋認証装置による家畜管理システム
US20040119831A1 (en) * 1999-12-03 2004-06-24 Yutaka Miyawaki Method of management of pictures and method of recording pictures of animals
US6845382B2 (en) 2000-12-22 2005-01-18 Louis B. Meadows Pet identification system and method
KR20020066771A (ko) 2001-02-13 2002-08-21 박현종 네트워크 기반의 개의 비문 정보를 이용한 신원확인 시스템
KR100414606B1 (ko) 2001-02-13 2004-01-07 박현종 네트워크 기반의 개의 비문 정보를 이용한 신원확인 시스템
USD458421S1 (en) 2001-08-23 2002-06-04 Dennis Jay Wasserman Pet identification tag
US6568354B1 (en) 2002-02-05 2003-05-27 Dennis J. Wasserman Pet identification tag with electronic display
JP4190209B2 (ja) 2002-05-23 2008-12-03 サージミヤワキ株式会社 鼻紋の照合方法
JP2003346148A (ja) 2002-05-23 2003-12-05 Surge Miyawaki Co Ltd 鼻紋の照合方法
KR20040008072A (ko) 2002-07-15 2004-01-28 김윤옥 소 개체관리 휴대형 정보 단말기 및 비문인식방법
KR20040006822A (ko) 2002-07-15 2004-01-24 김윤옥 인터넷 망과 네트워크 망을 통한 소 개체관리 시스템 및제어방법
KR100479937B1 (ko) 2002-07-15 2005-03-30 김윤옥 인터넷 망과 cdma 네트워크 망을 통한 소 개체 관리방법
KR100485922B1 (ko) 2002-07-15 2005-04-29 김윤옥 소 개체관리 휴대형 정보 단말기 및 비문인식방법
USD494326S1 (en) 2002-12-23 2004-08-10 Angela Rae Long Personalized animal/pet identification tag
USD491323S1 (en) 2003-04-14 2004-06-08 Amy Wilcox Pet identification collar
KR200337597Y1 (ko) 2003-10-20 2004-01-03 김주경 동물의 유동을 방지하는 유동방지기구
US7230539B2 (en) 2004-10-26 2007-06-12 Klein Elliot S Pet identification tag with internal cavity for transponder capsule
JP2007135501A (ja) 2005-11-21 2007-06-07 Atom System:Kk 鼻特徴情報生成装置及び鼻特徴情報生成プログラム
US20160110383A1 (en) * 2011-07-11 2016-04-21 Snoutscan Llc System and method for animal identification
US20150078626A1 (en) * 2013-09-17 2015-03-19 William Brian Kinard Animal / pet identification system and method based on biometrics

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
PCT International Search Report for PCT Counterpart Application No. PCT/KR2014/004487, 7 pp. (including English translation), (dated Oct. 27, 2014).
PCT Notification Concerning Transmittal of International Preliminary Report on Patentability (Chapter I of the Patent Cooperation Treaty) for PCT Counterpart Application No. PCT/KR2014/004487, 32 pp. (including English translation), (dated Dec. 3, 2015).
PCT Written Opinion of the International Searching Authority for PCT Counterpart Application No. PCT/KR2014/004487, 29 pp. (including English translation), (dated Oct. 27, 2014).

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200250733A1 (en) * 2019-02-01 2020-08-06 Societe Des Products Nestle S.A. Pet food recommendation devices and methods
US11538087B2 (en) * 2019-02-01 2022-12-27 Societe Des Produits Nestle Sa Pet food recommendation devices and methods

Also Published As

Publication number Publication date
EP4140294A1 (de) 2023-03-01
ES2936063T3 (es) 2023-03-14
EP3029603A4 (de) 2017-08-30
CA2925275C (en) 2022-11-08
EP3029603B1 (de) 2022-10-19
JP6466920B2 (ja) 2019-02-06
WO2014189250A2 (ko) 2014-11-27
JP2018191648A (ja) 2018-12-06
US10552673B2 (en) 2020-02-04
EP3029603A2 (de) 2016-06-08
KR101527801B1 (ko) 2015-06-11
JP6840703B2 (ja) 2021-03-10
KR20140137149A (ko) 2014-12-02
JP2019003669A (ja) 2019-01-10
US20190080160A1 (en) 2019-03-14
US20190080159A1 (en) 2019-03-14
US20160259970A1 (en) 2016-09-08
WO2014189250A3 (ko) 2015-01-15
CA2925275A1 (en) 2014-11-27
JP2016519941A (ja) 2016-07-11
US10565440B2 (en) 2020-02-18

Similar Documents

Publication Publication Date Title
US10565440B2 (en) Device for recognizing animal's identity by using animal nose prints
KR101494716B1 (ko) 동물들의 코무늬를 이용한 동물 개체 인식 장치
KR101494717B1 (ko) 동물들의 코무늬를 이용한 동물 개체 인식 장치
US9489560B2 (en) On-the go touchless fingerprint scanner
US11019250B2 (en) Method for implementing animal nose pattern biometric identification system on mobile devices
US10146993B2 (en) Non-invasive multimodal biometrical identification system of animals
CN107239772B (zh) 掌纹掌脉图像采集装置及通关闸机
US11450130B2 (en) Animal identification based on unique nose patterns
JP2000182050A (ja) 動物の個体識別装置
KR102325259B1 (ko) 반려동물 생애 관리 시스템 및 그 방법
US20210386035A1 (en) Animal identification using vision techniques
Kumar et al. Animal Biometrics
WO2015123374A1 (en) On-the-go touchless fingerprint scanner
JPH11161796A (ja) 動物の個体識別装置および画像処理方法
Gailey Computer systems for photo-identification and theodolite tracking of cetaceans
Cameron et al. A biometric approach to laboratory rodent identification
JP2000137811A (ja) 個体識別装置
WO2001064029A1 (en) Method and apparatus for livestock identification
KR20230001452U (ko) 비문이 미끄러지지 않도록 입력창 주위에 가이드를 구비한 비문 이미지 획득장치
CN117974749A (zh) 基于可见光与红外图像融合的鸡胸角测量方法、系统及介质
KR20240002171A (ko) 비문이 접촉하는 탄성접촉부를 구비한 비문 이미지 획득장치

Legal Events

Date Code Title Description
AS Assignment

Owner name: ISCILAB CORPORATION, KOREA, REPUBLIC OF

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:WEE, NAM SOOK;CHOI, SU JIN;KIM, HAENG MOON;REEL/FRAME:038688/0906

Effective date: 20160314

STPP Information on status: patent application and granting procedure in general

Free format text: PUBLICATIONS -- ISSUE FEE PAYMENT VERIFIED

STCF Information on status: patent grant

Free format text: PATENTED CASE

CC Certificate of correction
MAFP Maintenance fee payment

Free format text: PAYMENT OF MAINTENANCE FEE, 4TH YR, SMALL ENTITY (ORIGINAL EVENT CODE: M2551); ENTITY STATUS OF PATENT OWNER: SMALL ENTITY

Year of fee payment: 4